Give a coding agent a current, maintained reference for the LanceDB APIs.
A coding agent writing LanceDB code from memory writes whatever was in its
training data: often an older API, sometimes one that never existed. The
LanceDB agent plugin fixes that by supplying the current reference at the point
the agent needs it.
The plugin
It ships as the lancedb plugin in
the lancedb-agent-plugins
repository, which is also a plugin marketplace. Install it with your agent’s
plugin manager:
The plugins installer shown under
Other agents is a cross-tool option. It detects which agent CLIs are on your
PATH (Cursor, GitHub Copilot CLI, VS Code, Grok Build, and Kimi Code, as well
as Claude Code and Codex) and installs through each one’s native plugin system.The plugin supplements the agent’s training data with current LanceDB
instructions. To pick up later revisions, refresh the marketplace:
/plugin marketplace update lancedb
codex plugin marketplace upgrade
npx plugins add lancedb/lancedb-agent-plugins
Where the plugin gets installed
Each plugin manager clones the repository, reads its marketplace manifest to find
the lancedb plugin, and hands the plugin to the agent’s own plugin store rather
than copying files into your project. Claude Code and Codex keep it under
~/.claude/plugins and ~/.codex/plugins respectively, so the plugin is available
in every project on the machine.The npx plugins installer defaults to the same user-wide scope. Pass
--scope project to record the plugin in the current repository instead, so that
everyone working in it gets the same plugin, or -t <target> to install for a
single agent rather than every one it detects:
Not every agent supports project scope; npx plugins targets lists what it
found and what each target supports.
Automated prompt catalog
Explore the prompts covered by automated agent evaluations. Each prompt uses
example data and may require existing tables,
indexes, metadata, branches, version history, or running jobs. Replace the test
fixtures and connection details with your own before using a prompt.The evaluations run against a LanceDB Enterprise test database, with and
without the agent plugin. Evaluation coverage does not guarantee success for
every agent or model.Prompts are reproduced verbatim, including the answer formats used by automated
grading. Table names, local files, and placeholders in these prompts refer to
test fixtures. Supply your own data, database connection, and timestamps when
adapting them.70 prompts.
Tables and schema design
101: list tables by prefix
Some tables in this database came from a legacy load and are named with the prefix `eval_lt_`. Find every table whose name starts with `eval_lt_` and report them. This is a read-only check. End your answer with one line per table, each of exactly this form:TABLE: <table_name>
102: describe table
Give me an overview of the `eval_describe` table: its columns with their types, and how many rows it holds. This is a read-only check — do not modify anything. End your answer with one line per column, then two final lines, of exactly these forms. Write each `<type>` exactly as the API reports it, not as a synonym; for a nested type, write the outer type name with its element type and length in angle brackets (a 4-element vector of 8-bit ints would read `fixed_size_list<int8, 4>`):COLUMN: <column_name> = <type>ROWS: <row_count>COLUMNS: <number_of_columns>
103: table storage stats
How is the `eval_table_stats` table laid out in storage? Report its total size on disk and its fragment layout. This is a read-only check. End your answer with two lines of exactly this form:BYTES: <total_size_in_bytes>FRAGMENTS: <number_of_fragments>
104: create table with data
Create a new table called `eval_created_products` in this database with three columns: `sku` (string), `name` (string), and `price` (float64). Populate it with exactly these 4 rows:| sku | name | price ||---------|------------------|--------|| SKU-001 | Espresso Machine | 249.99 || SKU-002 | Milk Frother | 39.5 || SKU-003 | Coffee Grinder | 89.0 || SKU-004 | Digital Scale | 24.95 |
105: create empty table
Create a new, empty table called `eval_empty_events` in this database, ready for an event-logging pipeline to start writing into later. It needs three columns: `event_name` (string), `severity` (int64), and `payload` (string) — and no rows yet.
107: drop table
The `eval_drop_me` table was created by a test run and is no longer needed. Drop it from the database — and nothing else.
108: design tables from joined schema
My current DB has these 5 tables that join; how should I set up my tables in LanceDB?```products (product_id BIGINT PK, name TEXT, description TEXT, price NUMERIC(10,2), category_id BIGINT FK, brand_id BIGINT FK)categories (category_id BIGINT PK, name TEXT)brands (brand_id BIGINT PK, name TEXT, country TEXT)reviews (review_id BIGINT PK, product_id BIGINT FK, rating SMALLINT, body TEXT)order_items (order_id BIGINT, product_id BIGINT FK, quantity INT, unit_price NUMERIC(10,2))```In LanceDB the app will run one query: semantic search over product descriptions, filtered by category name and brand name, where every hit shows the product's name, price, category name, brand name, and average review rating. Order analytics stays in our warehouse. Description embeddings are 384-dimensional float32 vectors.Set the tables up in this database as you recommend: create them empty (schema only, no rows). Name the table the product search runs against `eval_schema_search`, and give any other table you create the `eval_schema_` prefix. Explain your design in your answer.
Data ingestion and updates
201: insert rows
Append these 5 new products to the `eval_insert_rows` table (the embedding columns can be left null):| id | title | category | price_usd | relevance_score ||-----|--------------|-----------|-----------|-----------------|| 100 | Clearance A | clearance | 9.99 | 0.1 || 101 | Clearance B | clearance | 10.99 | 0.1 || 102 | Clearance C | clearance | 11.99 | 0.1 || 103 | Clearance D | clearance | 12.99 | 0.1 || 104 | Clearance E | clearance | 13.99 | 0.1 |
202: delete rows by filter
All 'toys' products have been discontinued. Delete every row in category `toys` from the `eval_delete_rows` table — and nothing else.
203: delete all rows
The `eval_purge` table was loaded from a corrupted export and every row in it is suspect. Clear out all of its rows — but keep the table itself (and its schema) in place so it can be reloaded.
204: update rows by filter
Editorial flagged every 'books' product in the `eval_update_rows` table for re-review: set `relevance_score` to 0 for all rows in category `books`. Don't touch any other rows.
205: update with expression
Run a flash sale on the `eval_discount` table: cut the price of every product in category `electronics` by 10% — each one's `price_usd` becomes 90% of its current value. Leave every other category's prices alone.
206: upsert rows
Sync the `eval_upsert` table with this batch of refurbished products, matching on `id`: update the row if the id already exists, insert it if not — as one upsert operation, not separate update and insert passes. The embedding columns can be left null.| id | title | category | price_usd | relevance_score ||-----|-----------------------|----------|-----------|-----------------|| 2 | Refurbished Product 2 | refurb | 19.99 | 0.5 || 6 | Refurbished Product 6 | refurb | 29.99 | 0.5 || 200 | Refurb Special A | refurb | 15.0 | 0.5 || 201 | Refurb Special B | refurb | 18.0 | 0.5 |
207: add column backfill
Analytics needs integer cents. Add a `price_cents` integer column to the `eval_add_column` table, backfilled for every existing row as the price in cents (`price_usd` × 100, rounded to the nearest integer).
208: rename column
Rename the `price_usd` column of the `eval_rename_column` table to `price`, in place, keeping the data — do not rebuild or rewrite the table.
209: drop column
The legacy `image_clip_v1` embedding column of the `eval_drop_column` table is superseded and wasting storage. Remove that column — keep everything else.
210: cast column type
Downstream tooling needs 64-bit floats: change the type of the `relevance_score` column in the `eval_cast_column` table from float32 to float64, in place, keeping the values.
211: performant batch ingest
Make sure my ingest pattern will be performant. My ingest script is `ingest.py` in your working directory; it creates the table `eval_ingest_events` in this database and loads the day's 1,000 events into it. Fix whatever would make the ingest slow or hurt the table, keeping the same table name, schema, and rows, then run it against this database so the table ends up holding exactly those 1,000 events.
Queries and search
301: filtered count
How many products in the `eval_filter_count` table are in category `books` with a price above $100? This is a read-only check. End your answer with a line of exactly this form:COUNT: <n>
302: filtered ids
Which products in the `eval_filter_ids` table are in category `electronics` and priced under $50? This is a read-only check. End your answer with one line per matching product, each of exactly this form:ID: <id>
303: lookup row by id
Look up the product with id 12 in the `eval_lookup_row` table and report its title. This is a read-only check. End your answer with a line of exactly this form:TITLE: <title>
304: top k cheapest
What are the 3 cheapest products in the `eval_cheapest` table? This is a read-only check. End your answer with a line of exactly this form, ids ordered from cheapest up:CHEAPEST: <id>,<id>,<id>
305: paginated read
Our storefront lists the `eval_pagination` table's products cheapest-first, 3 to a page. A customer is looking at page 2 — which products do they see? This is a read-only check. End your answer with a line of exactly this form, the three ids in display (price-ascending) order:PAGE2: <id>,<id>,<id>
306: pick column for image search
I have a LanceDB table called `eval_pick_column`. I want to let users upload a photo and find visually similar products. Which column should the similarity search run against? If there's more than one candidate, tell me which one to use and why. This is a read-only check — do not modify the table, its rows, or its metadata. End your answer with a line of exactly this form, naming the single column to use:RECOMMENDATION: <column_name>
307: vector search similar
A shopper is viewing product 5 in the `eval_vector_search` table and we want a 'similar items' shelf. Using product 5's `image_clip_v2` embedding as the query, find the single most visually similar OTHER product. This is a read-only check. End your answer with a line of exactly this form:MATCH: <id>
308: vector search filtered
Same 'similar items' feature, but constrained to a department: using product 8's `image_clip_v2` embedding as the query on the `eval_vector_filtered` table, find the single most similar product in category `books` (product 8 itself is not in that category). This is a read-only check. End your answer with a line of exactly this form:MATCH: <id>
309: full text search
A customer remembers a product description mentioning 'waterproof'. Using full-text search on the `eval_fts_search` table, find which product that is. This is a read-only check. End your answer with a line of exactly this form:MATCH: <id>
310: fts multi match
Marketing wants a list of every product in the `eval_fts_multi` table whose description mentions 'wireless'. Use full-text search to find them all. This is a read-only check. End your answer with one line per matching product, each of exactly this form:ID: <id>
311: hybrid search
Find the product in the `eval_hybrid_search` table most relevant to a shopper who typed the keyword 'solar' and whose taste profile is product 4's `image_clip_v2` embedding — i.e. combine keyword relevance and vector similarity (a hybrid search). This is a read-only check. End your answer with a line of exactly this form:MATCH: <id>
312: explain query plan
Before we ship a price filter on the `eval_explain_plan` table, confirm from the server's query plan whether a filter on `price_usd` (for example `price_usd < 50`) is served by an index or by scanning the table. This is a read-only check — do not build or drop anything. End your answer with exactly one of these lines:PLAN: indexPLAN: scan
Index management
401: build index
Queries that filter the `eval_build_index` table by `category` are slow. Build the appropriate index to speed them up.
402: create vector index
Vector similarity queries against the `image_clip_v2` column of the `eval_vector_index` table are running as brute-force scans. Build an ANN index on that column using cosine distance, and make sure it finishes building.
403: create fts index
Enable keyword (full-text) search over the `body_text` column of the `eval_fts_index` table, and make sure the index finishes building.
404: inspect indexes
Audit the indexes on the `eval_index_report` table: what indexes exist, and how many rows does the index on `category` cover? This is a read-only check. End your answer with one line per index plus one final line, of exactly these forms:INDEX: <index_name>INDEXED_ROWS: <rows covered by the category index>
405: index coverage
The scalar index on `price_usd` of the `eval_index_coverage` table was built a while ago, and data has landed in the table since. Check the index's coverage: how many rows does it currently cover, and how many rows are not yet indexed? This is a read-only check — do not rebuild or modify anything. End your answer with two lines of exactly this form:INDEXED: <rows covered by the index>UNINDEXED: <rows not yet covered>
406: drop index
The index on `price_usd` in the `eval_drop_index` table was an experiment and is slowing down writes. Drop that index.
Version history and tags
501: version history
How many versions does the `eval_versions` table have? This is a read-only check — do not create new versions. End your answer with a line of exactly this form:VERSIONS: <n>
502: time travel read
Without modifying the `eval_timetravel` table, compare it to its past: how many rows did it have at version 1, and how many does it have now? End your answer with two lines of exactly this form:THEN: <rows at version 1>NOW: <rows now>
503: restore version
Someone ran a bad cleanup against the `eval_restore` table and rows went missing. Roll the table back to the version immediately before the deletion, so the missing rows are back.
504: tag version
We're about to run a risky migration on the `eval_tag_create` table. Tag its current version as `prod-baseline` so we can refer back to it later.
505: read at tag
The `eval_tag_read` table's version tagged `release-v1` is what we shipped. Without modifying the table, compare then and now: how many rows did it have at the `release-v1` tag, and how many does it have now? End your answer with two lines of exactly this form:TAGGED: <rows at the release-v1 tag>NOW: <rows now>
506: move tag
The `latest` tag on the `eval_tag_move` table still points at an old version, but the table has moved on since. Update the `latest` tag so it points at the table's current version. Leave any other tags alone.
507: delete tag
The `deprecated` tag on the `eval_tag_delete` table is obsolete. Remove that tag — leave any other tags alone.
508: compare two versions
Compare version 3 and version 7 of the `eval_history_compare` table and explain, in plain English, what happened to it in between — cover the schema, the indexes, and the data. This is a read-only investigation: do not modify the table.Report only what changed AFTER version 3, up to and including version 7 — version 3 is the starting point, so whatever happened at version 3 itself is not one of the changes.End your answer with three lines of exactly these forms. The first two are comma-separated lists of column names (or NONE): the columns added in that range, and the columns that gained an index in that range. The third is exactly one of the five words offered, describing the row-level data changes in that range:COLUMNS_ADDED: <column_name>[, <column_name>...]INDEXES_ADDED: <column_name>[, <column_name>...]DATA_CHANGES: APPEND_THEN_DELETE | DELETE_THEN_APPEND | APPEND_ONLY | DELETE_ONLY | NONE
509: changes since date
Our last schema review of the `eval_history_since` table signed off at {cutoff}. Did any schema or index changes happen to that table since then? Explain in plain English what you find, and say which changes were not schema or index changes so we know they were considered. This is a read-only investigation: do not modify the table.This table also has columns and indexes that predate that timestamp — those are not what the review is asking about, so report only what changed at or after it.End your answer with four lines of exactly these forms: every table version at or after that timestamp that changed the schema or an index — version numbers only, comma-separated, ascending (or NONE); the column added by a schema change at or after that timestamp (or NONE); the column covered by an index built at or after that timestamp (or NONE); and every table version at or after that timestamp whose change was NOT a schema or index change — version numbers only, comma-separated, ascending (or NONE):CHANGED_VERSIONS: <version>[, <version>...]COLUMN_ADDED: <column_name>INDEX_ADDED: <column_name>OTHER_CHANGES: <version>[, <version>...]
510: trace column to job
Someone added a column called `foo` to the `eval_history_column_job_{db}` table and nobody wrote down why. Find the exact version where `foo` appeared, then find the background job that was pinned to that same table version and report what it is doing and its current state. There are several jobs on this table, pinned to different versions of it, so make sure the one you report is the one pinned to that version. This is a read-only investigation: do not cancel or modify anything.End your answer with four lines of exactly these forms (full job id, no truncation). Write `<job_type>` and `<current_state>` exactly as the API reports them, not as synonyms or prose:VERSION: <version_number>JOB: <job_id>TYPE: <job_type>STATE: <current_state>
511: compact and clean versions
Our table has grown to 3 billion rows and 400 versions after many tiny appends. Reads are getting slower and storage keeps growing. How would you improve this without breaking the readers?The table `eval_maintenance_{db}` in this database is a small copy built from many small appends. Imagine it at the production scale above with the same write pattern. Inspect it and apply the changes you would make to the real table. Preserve all current data. Two readers depend on it: the app reads the latest version, and a month-end report reads the version tagged `month-end`.
Column metadata and lineage
601: add all metadata types
Document the two image-embedding columns in the `eval_add_metadata` table: write a clear description for each, tag each with its field type, the model that produced it, and its version, mark the newer one as the latest, and group both under a single logical column.
602: delete one metadata key
The `model` tag on the `image_clip_v1` column of `eval_delete_metadata` is wrong. Remove just that tag — leave the column's description and other tags alone.
603: explain column origin
Some columns in the `eval_lin_origin` table were generated by the platform rather than loaded with the data. Investigate the column `title_sentiment`: is it produced by a UDF or by a SQL expression, what produces it, and which columns does it read? This is a read-only investigation: do not modify the table.End your answer with lines of exactly these forms — one INPUT line per input column. Write the generator exactly as the table's own metadata records it (the UDF's registered name, or the SQL expression), not a paraphrase:KIND: <udf or expression>GENERATOR: <udf name or SQL expression>INPUT: <input_column_name>
604: find downstream dependents
We want to drop the `price_usd` column from the `eval_lin_deps` table, but some of the table's generated columns are derived from other columns. Find every column that depends on `price_usd` — directly, or indirectly through another generated column. This is a read-only investigation: do not modify the table.End your answer with one line per dependent column, each of exactly this form (do not list `price_usd` itself):DEPENDENT: <column_name>
605: trace column refresh job
The generated column `discounted_price` in the `eval_lin_job_{db}` table is filled in by a background refresh job, and we need to know which job that is. There is more than one job on this table (a long-lived database can also hold finished or cancelled leftovers), so make sure the one you report is the currently-running job that refreshes `discounted_price`. This is a read-only investigation: do not cancel or modify anything.End your answer with three lines of exactly these forms (full job id, no truncation). Write `<job_type>` and `<current_state>` exactly as the API reports them, not as synonyms or prose:JOB: <job_id>TYPE: <job_type>STATE: <current_state>
Namespaces
701: create namespace
Create a namespace called `analytics` in this database, ready for new tables to be organized under it.
702: create table in namespace
Under the `staging` namespace of this database, create a table called `sessions` with columns `session_id` (string) and `duration_s` (float64), holding exactly these 2 rows:| session_id | duration_s ||------------|------------|| a1b2 | 12.5 || c3d4 | 3.25 |
703: list namespace tables
What tables live inside the `warehouse` namespace of this database? This is a read-only check. End your answer with one line per table, each of exactly this form:TABLE: <table_name>
704: query namespace table
Our SRE dashboards read from the `ops` namespace of this database, which holds a `latency` table of per-service p99 latency readings. Which service has the highest `p99_ms`, and what is that value? This is a read-only check. End your answer with two lines of exactly this form:SERVICE: <service_name>P99: <p99_ms value>
Table branches
801: create branch
Create a branch called `experiment-reindex` on the `eval_create_branch` table, off its current state.
802: switch branch and write
Switch to the `wip-branch` branch of the `eval_switch_branch` table and add a description to its `category` column. The change should land on `wip-branch`, not on main.
803: cherry pick branch
The `ready-to-cherry-pick` branch of the `eval_cherry_pick` table has been reviewed and approved. Cherry-pick it onto main so its changes land on the main table.
804: delete branch
We no longer need the `stale-2024` branch on the `eval_delete_branch` table. Delete it.
Background jobs
901: list running jobs
What jobs are currently running against the `eval_job_list_{db}` table? Give each one's type and current state. This is a read-only check — do not cancel or modify anything. End your answer with one line per currently-running job, each of exactly this form (full job id, no truncation). Write `<job_type>` and `<current_state>` exactly as the API reports them, not as synonyms or prose — a compaction job that had failed would read `RUNNING: j1-xy-0ab-cdef | table_compaction | failed`:RUNNING: <job_id> | <job_type> | <current_state>
902: describe job
An automated maintenance job was recently started against the `eval_job_describe_{db}` table. Find it and report what kind of job it is and its current status. Do not cancel or modify it. End your answer with lines of exactly these forms (full job id, no truncation). Write the type and state exactly as the API reports them, not as synonyms or prose — a compaction job that had failed would read `TYPE: table_compaction` and `STATE: failed`:JOB: <job_id>TYPE: <job_type>STATE: <current_state>
903: cancel job
A cache-prewarm job against the `eval_job_cancel_{db}` table is hogging resources. Cancel it, and leave every other job alone. End your answer with a line of exactly this form (the full id of the job you cancelled, no truncation):CANCELLED: <job_id>
Multi-step workflows
1001: create embed search
Set up a small semantic-search demo from scratch in this database. Create a new table called `eval_embed_search` with exactly three columns: `id` (int64), `text` (string), and `embedding` — a fixed-size vector of floats holding an embedding of `text`, computed with the `all-MiniLM-L6-v2` sentence-transformers model (embed each `text` exactly as written; use that same model for every row and for the query below). Populate it with exactly these 8 rows:| id | text ||----|------|| 1 | Espresso machine with a 15-bar pump for rich, crema-topped coffee || 2 | Burr coffee grinder for fresh, evenly ground espresso beans || 3 | Pour-over coffee kettle with a gooseneck spout for slow brewing || 4 | Reusable coffee filter for drip brewers and pour-over cones || 5 | Waterproof hiking boots for rocky mountain trails || 6 | Lightweight hiking backpack with a hydration sleeve for day trails || 7 | Trekking poles for steep mountain trail descents || 8 | Merino wool hiking socks for long days on the trail |Then run two vector searches against the table, both using cosine distance: (1) a 'similar items' search using row 2's own stored embedding as the query, returning the 3 nearest OTHER rows; (2) a text search — embed the query `gear for a mountain hike` with the same model and return the 3 nearest rows. End your answer with two lines of exactly this form, ids comma-separated, nearest first:NEAREST: <id>,<id>,<id>QUERY_HITS: <id>,<id>,<id>
1002: build all indexes
The `eval_index_all` table is about to back our product search page. The page will run: visual similarity search on the `image_clip_v2` embeddings (cosine distance), keyword search over `body_text`, exact-match filters on `category`, and range filters on `price_usd` — often combined with the searches. Today the table has no indexes at all. Build all the indexes this table will need to make search fast, and make sure every one of them finishes building. Don't build indexes the page won't use — each one costs build time and slows writes — and note that `image_clip_v1` is a legacy column nothing queries anymore. Do not modify the table's rows or schema.
1003: embedding models in branches
We're choosing an embedding model for the `body_text` column of the `eval_embed_branches` table (24 product descriptions) and want to compare 3 candidates side by side without touching the production data. The candidates: `all-MiniLM-L6-v2`, `bge-small-en-v1.5`, and `gte-small`. Run the experiment on branches: for each candidate, create a branch of the table off main and, on that branch only, add a fixed-size float vector column called `body_embedding` holding that model's embedding of every row's `body_text` (embed the text exactly as stored). Then run the same search on each branch, using cosine distance — the 3 products most similar to the query `a paperback novel to read on the beach`, embedded with that branch's model — so we can compare what each model returns. Main must be left exactly as it is: no new column, no new rows, no other writes. End your answer with one line per branch, each of exactly this form (the model named the way its provider does, e.g. `all-MiniLM-L6-v2`; the three ids comma-separated, most similar first):BRANCH: <branch_name> | <embedding_model> | <id>,<id>,<id>
1004: parquet to lance embed
Convert this Parquet dataset to Lance and add a vector column from the description field. The dataset is `products.parquet` in your working directory. Load it into a new table called `eval_parquet_products` in this database, keeping every Parquet column and row exactly as they are, and add a column `description_embedding`: a fixed-size vector of floats holding an embedding of each row's `description`, computed with a text-embedding model of your choice (the same model for every row).
1005: rag chatbot citations
Build a RAG chatbot over this table with LangChain. Return citations.The table is `eval_rag_docs` in this database: help-center articles with `doc_id`, `title`, and `body`. The chatbot gets no LLM API credentials, so generate its answers with a model that runs locally, or answer extractively from the retrieved passages. When the chatbot is working, ask it: "How long do I have to return a damaged item for a refund?" In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. Include the chatbot's prose answer above the result lines for manual inspection. End your answer with two lines of exactly this form. ANSWER must contain only the number of days as decimal digits, with no units or other text:ANSWER: <number of days>CITATIONS: <doc_id>,<doc_id>,... (the articles the chatbot cited)
1006: fastapi similar endpoint
Add a recommend-similar-products endpoint to this FastAPI service.The service is `app.py` in your working directory; it serves the `eval_rec_products` table in this database. Add an endpoint that returns the products most similar to a given product, using the `image_clip_v2` vectors with either cosine or L2 distance and excluding the query product itself. Keep the existing endpoint working. Then run the service and call your endpoint for product 5, asking for 3 results. In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. End your answer with a line of exactly this form, ids nearest first:SIMILAR: <id>,<id>,<id>
1007: agent tool past conversations
Give my agent a tool that queries this table for relevant past conversations.The table is `eval_conversations` in this database, one row per past support conversation (`conversation_id`, `user_id`, `summary`). The agent serves one signed-in user at a time. Build the tool, then call it the way the agent would for user `u-ana` asking "my card was charged twice". In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. End your answer with a line of exactly this form, listing the conversation ids your tool returned:CONVERSATIONS: <id>,<id>,...
Was this page helpful?
{"docsConfig":{"theme":"mint","$schema":"https://mintlify.com/docs.json","name":"LanceDB","colors":{"primary":"#FF6B35","light":"#FF8A5C","dark":"#E55A2B"},"logo":{"light":"https://mintcdn.com/lancedb-bcbb4faf/6L0IRVkfdlgMU1Pw/static/assets/logo/dark-lancedb-logo.svg?fit=max\u0026auto=format\u0026n=6L0IRVkfdlgMU1Pw\u0026q=85\u0026s=d0592e59a29df2477080a857e13e341f","dark":"https://mintcdn.com/lancedb-bcbb4faf/t0AGunBr6LulGkIJ/static/assets/logo/light-lancedb-logo.svg?fit=max\u0026auto=format\u0026n=t0AGunBr6LulGkIJ\u0026q=85\u0026s=9b508fcb8d83cbc9bd05e57227d739e9","href":"https://lancedb.com"},"favicon":"https://mintcdn.com/lancedb-bcbb4faf/0sS6vrpmM3KSVyss/static/favicon.ico?fit=max\u0026auto=format\u0026n=0sS6vrpmM3KSVyss\u0026q=85\u0026s=84b9ea3dfc312b7a4dfc99219e229b33","appearance":{"strict":false},"background":{"color":{"light":"#FAF5F0","dark":"#000000"}},"navbar":{"links":[{"href":"https://www.lancedb.com/contact","label":"Support"}]},"footer":{"socials":{"github":"https://github.com/lancedb","linkedin":"https://www.linkedin.com/company/lancedb","x":"https://x.com/lancedb","discord":"https://discord.gg/AUEWnJ7Txb"},"links":[{"items":[{"label":"Cookie settings","href":"#cookie-settings"}]}]},"fonts":{"family":"Inter"},"styling":{"codeblocks":{"theme":{"light":"vitesse-light","dark":"catppuccin-mocha"}},"latex":true},"integrations":{"ga4":{"measurementId":"G-1TMC2PR69E"}},"banner":{"content":"[AI research has outgrown the modern data stack. We raised $57.5M to rebuild it. Read our Series B story →](https://www.lancedb.com/blog/lancedb-series-b-funding?utm_campaign=series-b-funding\u0026utm_content=docs\u0026utm_medium=lancedb\u0026utm_source=banner)","dismissible":true},"navigation":{"pages":[]}},"deployment":{"deploymentMetadataValue":{"subdomain":"lancedb-bcbb4faf","hasAudioNarration":false,"actualSubdomain":"lancedb-bcbb4faf","customDomains":["docs.lancedb.com"],"gitSource":{"type":"github","owner":"lancedb","repo":"docs","deployBranch":"assembled","contentDirectory":"docs","isPrivate":true,"mountPath":"","label":"","_id":"6af512fe0902081a53d8698c"},"feedback":{"thumbs":true,"edits":false,"issues":false},"entitlements":{"AI_CHAT":{"status":"ENABLED"},"WEB_MCP":{"status":"ENABLED"},"NARRATION":{"status":"ENABLED"}},"buildId":"6b008022286e5a483262aeff","clientVersion":"0.0.4227","searchFilterCounts":[{"language":"default","tag":"Documentation","count":366},{"language":"default","tag":"Enterprise","count":343},{"language":"default","tag":"Embedded","count":340},{"language":"default","tag":"API Reference","count":113},{"language":"default","tag":"REST API","count":110},{"language":"default","tag":"Feature Engineering","count":90},{"language":"default","tag":"Table","count":88},{"language":"default","tag":"Integrations","count":84},{"language":"default","tag":"SQL Reference","count":80},{"language":"default","tag":"Geneva","count":68}]},"assistantConfig":{"enableAskAiButton":false,"enableStarterQuestions":false,"enableAskAiSearchBar":true},"basePath":"","hasChatPermissions":true,"hasWebMcpPermissions":true,"toggles":[{"name":"kill-isr","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-rbac-roles","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-mintlify-slack","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-integrations-settings","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"rbac-expansion","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-pr-grouping","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-prefix","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"analytics-v7","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"user-flows-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"no-publish-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-onboarding","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving-live","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"stripTemperatureMiddleware","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"support-platform-pylon-zendesk","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"instant-rollback","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications-server","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-fable","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fixed-ai-action-pricing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"site-roles","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automation-resolve-merge-conflict","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"scoped-publishing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"mintql","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"page-level-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-version-history-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"billing-page-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"full-page-search","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workspace-projects","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"assistant-v5","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fork-pr-previews","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"revalidation-jwt","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-sso-scim","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"hobby-abuse-scan","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"dashboard-inbox","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-signup","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-dual-write","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-enabled","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"approve-review-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-metadata","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"sso-setup-deadline-banner","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-serve","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"live-agents","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"starter-pricing","enabled":true,"variant":{"name":"control","enabled":true,"payload":{"type":"string","value":"control"},"feature_enabled":true},"impressionData":false},{"name":"llms-txt-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"wildcard-email-auth","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false}],"googleFontsInlined":false},"customCss":[{"filePath":"static/styles/style.css","content":":root{--banner-left: #e4d8f8;--banner-right: #e55a2b}@font-face{font-family:\"LanceDB KaTeX AMS\";src:url(\"https://cdn.jsdelivr.net/npm/katex@0.16.0/dist/fonts/KaTeX_AMS-Regular.woff2\") format(\"woff2\");font-display:swap;font-style:normal;font-weight:400}.katex .mathbb,.katex .textbb{font-family:\"LanceDB KaTeX AMS\", \"KaTeX_AMS\", serif!important}#banner,:where(.banner,[data-banner],[class*=\"Banner_banner\"]){background:linear-gradient(90deg, var(--banner-left) 0%, var(--banner-right) 100%)!important;color:#1b0f09!important;border:none!important}#banner a,:where(.banner,[data-banner],[class*=\"Banner_banner\"]) a{color:#1b0f09!important;font-size:0.9rem;font-weight:500;text-decoration-color:rgba(27, 15, 9, 0.35)}#banner button,:where(.banner,[data-banner],[class*=\"Banner_banner\"]) button{color:#1b0f09!important}"}]}
{"page":{"pagePath":"enterprise/v1-13-0/basics/agentic-skills","slug":"/enterprise/v1-13-0/basics/agentic-skills","pageMetadata":{"title":"Agentic skills","sidebarTitle":"Agentic Skills","description":"Give a coding agent a current, maintained reference for the LanceDB APIs.","icon":"robot","keywords":["agents","plugin","claude code","codex","cursor","copilot"],"href":"/enterprise/v1-13-0/basics/agentic-skills"},"description":{"compiledSource":"\"use strict\";\nconst {jsx: _jsx} = arguments[0];\nconst {useMDXComponents: _provideComponents} = arguments[0];\nfunction _createMdxContent(props) {\n const _components = {\n p: \"p\",\n ..._provideComponents(),\n ...props.components\n };\n return _jsx(_components.p, {\n children: \"Give a coding agent a current, maintained reference for the LanceDB APIs.\"\n });\n}\nfunction MDXContent(props = {}) {\n const {wrapper: MDXLayout} = {\n ..._provideComponents(),\n ...props.components\n };\n return MDXLayout ? _jsx(MDXLayout, {\n ...props,\n children: _jsx(_createMdxContent, {\n ...props\n })\n }) : _createMdxContent(props);\n}\nreturn {\n default: MDXContent\n};\n","frontmatter":{},"scope":{}},"pageType":"default","mdx":{"mdxSource":{"compiledSource":"\"use strict\";const{Fragment:_Fragment,jsx:_jsx,jsxs:_jsxs}=arguments[0];const{useMDXComponents:_provideComponents}=arguments[0];function _createMdxContent(props){const _components={a:`a`,code:`code`,p:`p`,pre:`pre`,span:`span`,strong:`strong`,..._provideComponents(),...props.components},{Accordion,AccordionGroup,CodeBlock,CodeGroup,Heading}=_components;if(!Accordion)_missingMdxReference(`Accordion`,true);if(!AccordionGroup)_missingMdxReference(`AccordionGroup`,true);if(!CodeBlock)_missingMdxReference(`CodeBlock`,true);if(!CodeGroup)_missingMdxReference(`CodeGroup`,true);if(!Heading)_missingMdxReference(`Heading`,true);return _jsxs(_Fragment,{children:[_jsx(_components.p,{children:`A coding agent writing LanceDB code from memory writes whatever was in its\ntraining data: often an older API, sometimes one that never existed. The\nLanceDB agent plugin fixes that by supplying the current reference at the point\nthe agent needs it.`}),`\n`,_jsx(Heading,{level:`2`,id:`plugin`,children:`The plugin`}),`\n`,_jsxs(_components.p,{children:[`It ships as the `,_jsx(_components.code,{children:`lancedb`}),` plugin in\nthe `,_jsx(_components.a,{href:`https://github.com/lancedb/lancedb-agent-plugins`,children:`lancedb-agent-plugins`}),`\nrepository, which is also a plugin marketplace. Install it with your agent’s\nplugin manager:`]}),`\n`,_jsxs(CodeGroup,{children:[_jsx(CodeBlock,{filename:`Claude Code`,numberOfLines:`3`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`3`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`/plugin marketplace add lancedb/lancedb-agent-plugins`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`/plugin install lancedb@lancedb`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),_jsx(CodeBlock,{filename:`Codex`,icon:`terminal`,numberOfLines:`2`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`2`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`codex`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugin`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` marketplace`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` add`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` lancedb/lancedb-agent-plugins`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`codex`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugin`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` add`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` lancedb@lancedb`})]}),`\n`]})})}),_jsx(CodeBlock,{filename:`Other agents`,icon:`terminal`,numberOfLines:`1`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`1`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`npx`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugins`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` add`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` lancedb/lancedb-agent-plugins`})]}),`\n`]})})})]}),`\n`,_jsxs(_components.p,{children:[`The `,_jsx(_components.a,{href:`https://www.npmjs.com/package/plugins`,children:_jsx(_components.code,{children:`plugins`})}),` installer shown under\n`,_jsx(_components.strong,{children:`Other agents`}),` is a cross-tool option. It detects which agent CLIs are on your\n`,_jsx(_components.code,{children:`PATH`}),` (Cursor, GitHub Copilot CLI, VS Code, Grok Build, and Kimi Code, as well\nas Claude Code and Codex) and installs through each one’s native plugin system.`]}),`\n`,_jsx(_components.p,{children:`The plugin supplements the agent’s training data with current LanceDB\ninstructions. To pick up later revisions, refresh the marketplace:`}),`\n`,_jsxs(CodeGroup,{children:[_jsx(CodeBlock,{filename:`Claude Code`,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`/plugin marketplace update lancedb`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),_jsx(CodeBlock,{filename:`Codex`,icon:`terminal`,numberOfLines:`1`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`1`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`codex`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugin`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` marketplace`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` upgrade`})]}),`\n`]})})}),_jsx(CodeBlock,{filename:`Other agents`,icon:`terminal`,numberOfLines:`1`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`1`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`npx`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugins`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` add`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` lancedb/lancedb-agent-plugins`})]}),`\n`]})})})]}),`\n`,_jsxs(Accordion,{title:`Where the plugin gets installed`,id:`where-the-plugin-gets-installed`,children:[_jsxs(_components.p,{children:[`Each plugin manager clones the repository, reads its marketplace manifest to find\nthe `,_jsx(_components.code,{children:`lancedb`}),` plugin, and hands the plugin to the agent’s own plugin store rather\nthan copying files into your project. Claude Code and Codex keep it under\n`,_jsx(_components.code,{children:`~/.claude/plugins`}),` and `,_jsx(_components.code,{children:`~/.codex/plugins`}),` respectively, so the plugin is available\nin every project on the machine.`]}),_jsxs(_components.p,{children:[`The `,_jsx(_components.code,{children:`npx plugins`}),` installer defaults to the same user-wide scope. Pass\n`,_jsx(_components.code,{children:`--scope project`}),` to record the plugin in the current repository instead, so that\neveryone working in it gets the same plugin, or `,_jsx(_components.code,{children:`-t \u003ctarget>`}),` to install for a\nsingle agent rather than every one it detects:`]}),_jsx(CodeGroup,{children:_jsx(CodeBlock,{filename:`Bash`,icon:`terminal`,numberOfLines:`1`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`1`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#59873A`,\"--shiki-light-font-style\":`inherit`,\"--shiki-dark\":`#89B4FA`,\"--shiki-dark-font-style\":`italic`},children:`npx`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` plugins`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` add`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` lancedb/lancedb-agent-plugins`}),_jsx(_components.span,{style:{color:`#A65E2B`,\"--shiki-dark\":`#A6E3A1`},children:` --scope`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` project`}),_jsx(_components.span,{style:{color:`#A65E2B`,\"--shiki-dark\":`#A6E3A1`},children:` -t`}),_jsx(_components.span,{style:{color:`#B56959`,\"--shiki-dark\":`#A6E3A1`},children:` cursor`})]}),`\n`]})})})}),_jsxs(_components.p,{children:[`Not every agent supports project scope; `,_jsx(_components.code,{children:`npx plugins targets`}),` lists what it\nfound and what each target supports.`]})]}),`\n`,_jsx(Heading,{level:`2`,id:`automated-prompts`,children:`Automated prompt catalog`}),`\n`,_jsx(_components.p,{children:`Explore the prompts covered by automated agent evaluations. Each prompt uses\nexample data and may require existing tables,\nindexes, metadata, branches, version history, or running jobs. Replace the test\nfixtures and connection details with your own before using a prompt.`}),`\n`,_jsx(_components.p,{children:`The evaluations run against a LanceDB Enterprise test database, with and\nwithout the agent plugin. Evaluation coverage does not guarantee success for\nevery agent or model.`}),`\n`,_jsx(_components.p,{children:`Prompts are reproduced verbatim, including the answer formats used by automated\ngrading. Table names, local files, and placeholders in these prompts refer to\ntest fixtures. Supply your own data, database connection, and timestamps when\nadapting them.`}),`\n`,_jsx(_components.p,{children:`70 prompts.`}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-1`,children:`Tables and schema design`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`101: list tables by prefix`,id:`101-list-tables-by-prefix`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Some tables in this database came from a legacy load and are named with the prefix `eval_lt_`. Find every table whose name starts with `eval_lt_` and report them. This is a read-only check. End your answer with one line per table, each of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TABLE: \u003ctable_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`102: describe table`,id:`102-describe-table`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`6`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Give me an overview of the `eval_describe` table: its columns with their types, and how many rows it holds. This is a read-only check — do not modify anything. End your answer with one line per column, then two final lines, of exactly these forms. Write each `\u003ctype>` exactly as the API reports it, not as a synonym; for a nested type, write the outer type name with its element type and length in angle brackets (a 4-element vector of 8-bit ints would read `fixed_size_list\u003cint8, 4>`):\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`COLUMN: \u003ccolumn_name> = \u003ctype>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`ROWS: \u003crow_count>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`COLUMNS: \u003cnumber_of_columns>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`103: table storage stats`,id:`103-table-storage-stats`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"How is the `eval_table_stats` table laid out in storage? Report its total size on disk and its fragment layout. This is a read-only check. End your answer with two lines of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`BYTES: \u003ctotal_size_in_bytes>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`FRAGMENTS: \u003cnumber_of_fragments>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`104: create table with data`,id:`104-create-table-with-data`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`9`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`9`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Create a new table called `eval_created_products` in this database with three columns: `sku` (string), `name` (string), and `price` (float64). Populate it with exactly these 4 rows:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| sku | name | price |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`|---------|------------------|--------|`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| SKU-001 | Espresso Machine | 249.99 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| SKU-002 | Milk Frother | 39.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| SKU-003 | Coffee Grinder | 89.0 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| SKU-004 | Digital Scale | 24.95 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`105: create empty table`,id:`105-create-empty-table`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Create a new, empty table called `eval_empty_events` in this database, ready for an event-logging pipeline to start writing into later. It needs three columns: `event_name` (string), `severity` (int64), and `payload` (string) — and no rows yet.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`107: drop table`,id:`107-drop-table`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `eval_drop_me` table was created by a test run and is no longer needed. Drop it from the database — and nothing else.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`108: design tables from joined schema`,id:`108-design-tables-from-joined-schema`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`14`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`14`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`My current DB has these 5 tables that join; how should I set up my tables in LanceDB?`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"```\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`products (product_id BIGINT PK, name TEXT, description TEXT, price NUMERIC(10,2), category_id BIGINT FK, brand_id BIGINT FK)`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`categories (category_id BIGINT PK, name TEXT)`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`brands (brand_id BIGINT PK, name TEXT, country TEXT)`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`reviews (review_id BIGINT PK, product_id BIGINT FK, rating SMALLINT, body TEXT)`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`order_items (order_id BIGINT, product_id BIGINT FK, quantity INT, unit_price NUMERIC(10,2))`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"```\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`In LanceDB the app will run one query: semantic search over product descriptions, filtered by category name and brand name, where every hit shows the product's name, price, category name, brand name, and average review rating. Order analytics stays in our warehouse. Description embeddings are 384-dimensional float32 vectors.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Set the tables up in this database as you recommend: create them empty (schema only, no rows). Name the table the product search runs against `eval_schema_search`, and give any other table you create the `eval_schema_` prefix. Explain your design in your answer.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-2`,children:`Data ingestion and updates`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`201: insert rows`,id:`201-insert-rows`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`10`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`10`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Append these 5 new products to the `eval_insert_rows` table (the embedding columns can be left null):\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| id | title | category | price_usd | relevance_score |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`|-----|--------------|-----------|-----------|-----------------|`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 100 | Clearance A | clearance | 9.99 | 0.1 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 101 | Clearance B | clearance | 10.99 | 0.1 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 102 | Clearance C | clearance | 11.99 | 0.1 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 103 | Clearance D | clearance | 12.99 | 0.1 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 104 | Clearance E | clearance | 13.99 | 0.1 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`202: delete rows by filter`,id:`202-delete-rows-by-filter`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"All 'toys' products have been discontinued. Delete every row in category `toys` from the `eval_delete_rows` table — and nothing else.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`203: delete all rows`,id:`203-delete-all-rows`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `eval_purge` table was loaded from a corrupted export and every row in it is suspect. Clear out all of its rows — but keep the table itself (and its schema) in place so it can be reloaded.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`204: update rows by filter`,id:`204-update-rows-by-filter`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Editorial flagged every 'books' product in the `eval_update_rows` table for re-review: set `relevance_score` to 0 for all rows in category `books`. Don't touch any other rows.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`205: update with expression`,id:`205-update-with-expression`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Run a flash sale on the `eval_discount` table: cut the price of every product in category `electronics` by 10% — each one's `price_usd` becomes 90% of its current value. Leave every other category's prices alone.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`206: upsert rows`,id:`206-upsert-rows`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`9`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`9`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Sync the `eval_upsert` table with this batch of refurbished products, matching on `id`: update the row if the id already exists, insert it if not — as one upsert operation, not separate update and insert passes. The embedding columns can be left null.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| id | title | category | price_usd | relevance_score |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`|-----|-----------------------|----------|-----------|-----------------|`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 2 | Refurbished Product 2 | refurb | 19.99 | 0.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 6 | Refurbished Product 6 | refurb | 29.99 | 0.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 200 | Refurb Special A | refurb | 15.0 | 0.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 201 | Refurb Special B | refurb | 18.0 | 0.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`207: add column backfill`,id:`207-add-column-backfill`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Analytics needs integer cents. Add a `price_cents` integer column to the `eval_add_column` table, backfilled for every existing row as the price in cents (`price_usd` × 100, rounded to the nearest integer).\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`208: rename column`,id:`208-rename-column`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Rename the `price_usd` column of the `eval_rename_column` table to `price`, in place, keeping the data — do not rebuild or rewrite the table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`209: drop column`,id:`209-drop-column`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The legacy `image_clip_v1` embedding column of the `eval_drop_column` table is superseded and wasting storage. Remove that column — keep everything else.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`210: cast column type`,id:`210-cast-column-type`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Downstream tooling needs 64-bit floats: change the type of the `relevance_score` column in the `eval_cast_column` table from float32 to float64, in place, keeping the values.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`211: performant batch ingest`,id:`211-performant-batch-ingest`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Make sure my ingest pattern will be performant. My ingest script is `ingest.py` in your working directory; it creates the table `eval_ingest_events` in this database and loads the day's 1,000 events into it. Fix whatever would make the ingest slow or hurt the table, keeping the same table name, schema, and rows, then run it against this database so the table ends up holding exactly those 1,000 events.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-3`,children:`Queries and search`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`301: filtered count`,id:`301-filtered-count`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"How many products in the `eval_filter_count` table are in category `books` with a price above $100? This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`COUNT: \u003cn>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`302: filtered ids`,id:`302-filtered-ids`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Which products in the `eval_filter_ids` table are in category `electronics` and priced under $50? This is a read-only check. End your answer with one line per matching product, each of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`ID: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`303: lookup row by id`,id:`303-lookup-row-by-id`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Look up the product with id 12 in the `eval_lookup_row` table and report its title. This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TITLE: \u003ctitle>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`304: top k cheapest`,id:`304-top-k-cheapest`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"What are the 3 cheapest products in the `eval_cheapest` table? This is a read-only check. End your answer with a line of exactly this form, ids ordered from cheapest up:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`CHEAPEST: \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`305: paginated read`,id:`305-paginated-read`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Our storefront lists the `eval_pagination` table's products cheapest-first, 3 to a page. A customer is looking at page 2 — which products do they see? This is a read-only check. End your answer with a line of exactly this form, the three ids in display (price-ascending) order:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`PAGE2: \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`306: pick column for image search`,id:`306-pick-column-for-image-search`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"I have a LanceDB table called `eval_pick_column`. I want to let users upload a photo and find visually similar products. Which column should the similarity search run against? If there's more than one candidate, tell me which one to use and why. This is a read-only check — do not modify the table, its rows, or its metadata. End your answer with a line of exactly this form, naming the single column to use:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`RECOMMENDATION: \u003ccolumn_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`307: vector search similar`,id:`307-vector-search-similar`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"A shopper is viewing product 5 in the `eval_vector_search` table and we want a 'similar items' shelf. Using product 5's `image_clip_v2` embedding as the query, find the single most visually similar OTHER product. This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`MATCH: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`308: vector search filtered`,id:`308-vector-search-filtered`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Same 'similar items' feature, but constrained to a department: using product 8's `image_clip_v2` embedding as the query on the `eval_vector_filtered` table, find the single most similar product in category `books` (product 8 itself is not in that category). This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`MATCH: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`309: full text search`,id:`309-full-text-search`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"A customer remembers a product description mentioning 'waterproof'. Using full-text search on the `eval_fts_search` table, find which product that is. This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`MATCH: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`310: fts multi match`,id:`310-fts-multi-match`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Marketing wants a list of every product in the `eval_fts_multi` table whose description mentions 'wireless'. Use full-text search to find them all. This is a read-only check. End your answer with one line per matching product, each of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`ID: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`311: hybrid search`,id:`311-hybrid-search`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Find the product in the `eval_hybrid_search` table most relevant to a shopper who typed the keyword 'solar' and whose taste profile is product 4's `image_clip_v2` embedding — i.e. combine keyword relevance and vector similarity (a hybrid search). This is a read-only check. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`MATCH: \u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`312: explain query plan`,id:`312-explain-query-plan`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Before we ship a price filter on the `eval_explain_plan` table, confirm from the server's query plan whether a filter on `price_usd` (for example `price_usd \u003c 50`) is served by an index or by scanning the table. This is a read-only check — do not build or drop anything. End your answer with exactly one of these lines:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`PLAN: index`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`PLAN: scan`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-4`,children:`Index management`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`401: build index`,id:`401-build-index`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Queries that filter the `eval_build_index` table by `category` are slow. Build the appropriate index to speed them up.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`402: create vector index`,id:`402-create-vector-index`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Vector similarity queries against the `image_clip_v2` column of the `eval_vector_index` table are running as brute-force scans. Build an ANN index on that column using cosine distance, and make sure it finishes building.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`403: create fts index`,id:`403-create-fts-index`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Enable keyword (full-text) search over the `body_text` column of the `eval_fts_index` table, and make sure the index finishes building.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`404: inspect indexes`,id:`404-inspect-indexes`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Audit the indexes on the `eval_index_report` table: what indexes exist, and how many rows does the index on `category` cover? This is a read-only check. End your answer with one line per index plus one final line, of exactly these forms:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INDEX: \u003cindex_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INDEXED_ROWS: \u003crows covered by the category index>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`405: index coverage`,id:`405-index-coverage`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The scalar index on `price_usd` of the `eval_index_coverage` table was built a while ago, and data has landed in the table since. Check the index's coverage: how many rows does it currently cover, and how many rows are not yet indexed? This is a read-only check — do not rebuild or modify anything. End your answer with two lines of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INDEXED: \u003crows covered by the index>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`UNINDEXED: \u003crows not yet covered>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`406: drop index`,id:`406-drop-index`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The index on `price_usd` in the `eval_drop_index` table was an experiment and is slowing down writes. Drop that index.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-5`,children:`Version history and tags`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`501: version history`,id:`501-version-history`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"How many versions does the `eval_versions` table have? This is a read-only check — do not create new versions. End your answer with a line of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`VERSIONS: \u003cn>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`502: time travel read`,id:`502-time-travel-read`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Without modifying the `eval_timetravel` table, compare it to its past: how many rows did it have at version 1, and how many does it have now? End your answer with two lines of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`THEN: \u003crows at version 1>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`NOW: \u003crows now>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`503: restore version`,id:`503-restore-version`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Someone ran a bad cleanup against the `eval_restore` table and rows went missing. Roll the table back to the version immediately before the deletion, so the missing rows are back.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`504: tag version`,id:`504-tag-version`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"We're about to run a risky migration on the `eval_tag_create` table. Tag its current version as `prod-baseline` so we can refer back to it later.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`505: read at tag`,id:`505-read-at-tag`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `eval_tag_read` table's version tagged `release-v1` is what we shipped. Without modifying the table, compare then and now: how many rows did it have at the `release-v1` tag, and how many does it have now? End your answer with two lines of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TAGGED: \u003crows at the release-v1 tag>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`NOW: \u003crows now>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`506: move tag`,id:`506-move-tag`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `latest` tag on the `eval_tag_move` table still points at an old version, but the table has moved on since. Update the `latest` tag so it points at the table's current version. Leave any other tags alone.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`507: delete tag`,id:`507-delete-tag`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `deprecated` tag on the `eval_tag_delete` table is obsolete. Remove that tag — leave any other tags alone.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`508: compare two versions`,id:`508-compare-two-versions`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`10`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`10`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Compare version 3 and version 7 of the `eval_history_compare` table and explain, in plain English, what happened to it in between — cover the schema, the indexes, and the data. This is a read-only investigation: do not modify the table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Report only what changed AFTER version 3, up to and including version 7 — version 3 is the starting point, so whatever happened at version 3 itself is not one of the changes.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`End your answer with three lines of exactly these forms. The first two are comma-separated lists of column names (or NONE): the columns added in that range, and the columns that gained an index in that range. The third is exactly one of the five words offered, describing the row-level data changes in that range:`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`COLUMNS_ADDED: \u003ccolumn_name>[, \u003ccolumn_name>...]`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INDEXES_ADDED: \u003ccolumn_name>[, \u003ccolumn_name>...]`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`DATA_CHANGES: APPEND_THEN_DELETE | DELETE_THEN_APPEND | APPEND_ONLY | DELETE_ONLY | NONE`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`509: changes since date`,id:`509-changes-since-date`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`11`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`11`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Our last schema review of the `eval_history_since` table signed off at {cutoff}. Did any schema or index changes happen to that table since then? Explain in plain English what you find, and say which changes were not schema or index changes so we know they were considered. This is a read-only investigation: do not modify the table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`This table also has columns and indexes that predate that timestamp — those are not what the review is asking about, so report only what changed at or after it.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`End your answer with four lines of exactly these forms: every table version at or after that timestamp that changed the schema or an index — version numbers only, comma-separated, ascending (or NONE); the column added by a schema change at or after that timestamp (or NONE); the column covered by an index built at or after that timestamp (or NONE); and every table version at or after that timestamp whose change was NOT a schema or index change — version numbers only, comma-separated, ascending (or NONE):`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`CHANGED_VERSIONS: \u003cversion>[, \u003cversion>...]`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`COLUMN_ADDED: \u003ccolumn_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INDEX_ADDED: \u003ccolumn_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`OTHER_CHANGES: \u003cversion>[, \u003cversion>...]`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`510: trace column to job`,id:`510-trace-column-to-job`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`9`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`9`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Someone added a column called `foo` to the `eval_history_column_job_{db}` table and nobody wrote down why. Find the exact version where `foo` appeared, then find the background job that was pinned to that same table version and report what it is doing and its current state. There are several jobs on this table, pinned to different versions of it, so make sure the one you report is the one pinned to that version. This is a read-only investigation: do not cancel or modify anything.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"End your answer with four lines of exactly these forms (full job id, no truncation). Write `\u003cjob_type>` and `\u003ccurrent_state>` exactly as the API reports them, not as synonyms or prose:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`VERSION: \u003cversion_number>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`JOB: \u003cjob_id>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TYPE: \u003cjob_type>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`STATE: \u003ccurrent_state>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`511: compact and clean versions`,id:`511-compact-and-clean-versions`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Our table has grown to 3 billion rows and 400 versions after many tiny appends. Reads are getting slower and storage keeps growing. How would you improve this without breaking the readers?`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The table `eval_maintenance_{db}` in this database is a small copy built from many small appends. Imagine it at the production scale above with the same write pattern. Inspect it and apply the changes you would make to the real table. Preserve all current data. Two readers depend on it: the app reads the latest version, and a month-end report reads the version tagged `month-end`.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-6`,children:`Column metadata and lineage`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`601: add all metadata types`,id:`601-add-all-metadata-types`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Document the two image-embedding columns in the `eval_add_metadata` table: write a clear description for each, tag each with its field type, the model that produced it, and its version, mark the newer one as the latest, and group both under a single logical column.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`602: delete one metadata key`,id:`602-delete-one-metadata-key`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `model` tag on the `image_clip_v1` column of `eval_delete_metadata` is wrong. Remove just that tag — leave the column's description and other tags alone.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`603: explain column origin`,id:`603-explain-column-origin`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`8`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`8`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Some columns in the `eval_lin_origin` table were generated by the platform rather than loaded with the data. Investigate the column `title_sentiment`: is it produced by a UDF or by a SQL expression, what produces it, and which columns does it read? This is a read-only investigation: do not modify the table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`End your answer with lines of exactly these forms — one INPUT line per input column. Write the generator exactly as the table's own metadata records it (the UDF's registered name, or the SQL expression), not a paraphrase:`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`KIND: \u003cudf or expression>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`GENERATOR: \u003cudf name or SQL expression>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`INPUT: \u003cinput_column_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`604: find downstream dependents`,id:`604-find-downstream-dependents`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`6`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"We want to drop the `price_usd` column from the `eval_lin_deps` table, but some of the table's generated columns are derived from other columns. Find every column that depends on `price_usd` — directly, or indirectly through another generated column. This is a read-only investigation: do not modify the table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"End your answer with one line per dependent column, each of exactly this form (do not list `price_usd` itself):\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`DEPENDENT: \u003ccolumn_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`605: trace column refresh job`,id:`605-trace-column-refresh-job`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`8`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`8`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The generated column `discounted_price` in the `eval_lin_job_{db}` table is filled in by a background refresh job, and we need to know which job that is. There is more than one job on this table (a long-lived database can also hold finished or cancelled leftovers), so make sure the one you report is the currently-running job that refreshes `discounted_price`. This is a read-only investigation: do not cancel or modify anything.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"End your answer with three lines of exactly these forms (full job id, no truncation). Write `\u003cjob_type>` and `\u003ccurrent_state>` exactly as the API reports them, not as synonyms or prose:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`JOB: \u003cjob_id>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TYPE: \u003cjob_type>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`STATE: \u003ccurrent_state>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-7`,children:`Namespaces`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`701: create namespace`,id:`701-create-namespace`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Create a namespace called `analytics` in this database, ready for new tables to be organized under it.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`702: create table in namespace`,id:`702-create-table-in-namespace`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`7`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`7`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Under the `staging` namespace of this database, create a table called `sessions` with columns `session_id` (string) and `duration_s` (float64), holding exactly these 2 rows:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| session_id | duration_s |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`|------------|------------|`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| a1b2 | 12.5 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| c3d4 | 3.25 |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`703: list namespace tables`,id:`703-list-namespace-tables`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"What tables live inside the `warehouse` namespace of this database? This is a read-only check. End your answer with one line per table, each of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TABLE: \u003ctable_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`704: query namespace table`,id:`704-query-namespace-table`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`5`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Our SRE dashboards read from the `ops` namespace of this database, which holds a `latency` table of per-service p99 latency readings. Which service has the highest `p99_ms`, and what is that value? This is a read-only check. End your answer with two lines of exactly this form:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`SERVICE: \u003cservice_name>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`P99: \u003cp99_ms value>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-8`,children:`Table branches`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`801: create branch`,id:`801-create-branch`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Create a branch called `experiment-reindex` on the `eval_create_branch` table, off its current state.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`802: switch branch and write`,id:`802-switch-branch-and-write`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Switch to the `wip-branch` branch of the `eval_switch_branch` table and add a description to its `category` column. The change should land on `wip-branch`, not on main.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`803: cherry pick branch`,id:`803-cherry-pick-branch`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `ready-to-cherry-pick` branch of the `eval_cherry_pick` table has been reviewed and approved. Cherry-pick it onto main so its changes land on the main table.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`804: delete branch`,id:`804-delete-branch`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"We no longer need the `stale-2024` branch on the `eval_delete_branch` table. Delete it.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-9`,children:`Background jobs`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`901: list running jobs`,id:`901-list-running-jobs`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"What jobs are currently running against the `eval_job_list_{db}` table? Give each one's type and current state. This is a read-only check — do not cancel or modify anything. End your answer with one line per currently-running job, each of exactly this form (full job id, no truncation). Write `\u003cjob_type>` and `\u003ccurrent_state>` exactly as the API reports them, not as synonyms or prose — a compaction job that had failed would read `RUNNING: j1-xy-0ab-cdef | table_compaction | failed`:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`RUNNING: \u003cjob_id> | \u003cjob_type> | \u003ccurrent_state>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`902: describe job`,id:`902-describe-job`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`6`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"An automated maintenance job was recently started against the `eval_job_describe_{db}` table. Find it and report what kind of job it is and its current status. Do not cancel or modify it. End your answer with lines of exactly these forms (full job id, no truncation). Write the type and state exactly as the API reports them, not as synonyms or prose — a compaction job that had failed would read `TYPE: table_compaction` and `STATE: failed`:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`JOB: \u003cjob_id>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`TYPE: \u003cjob_type>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`STATE: \u003ccurrent_state>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`903: cancel job`,id:`903-cancel-job`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"A cache-prewarm job against the `eval_job_cancel_{db}` table is hogging resources. Cancel it, and leave every other job alone. End your answer with a line of exactly this form (the full id of the job you cancelled, no truncation):\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`CANCELLED: \u003cjob_id>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]}),`\n`,_jsx(Heading,{level:`3`,id:`agent-evals-10`,children:`Multi-step workflows`}),`\n`,_jsxs(AccordionGroup,{children:[_jsx(Accordion,{title:`1001: create embed search`,id:`1001-create-embed-search`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`18`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`18`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Set up a small semantic-search demo from scratch in this database. Create a new table called `eval_embed_search` with exactly three columns: `id` (int64), `text` (string), and `embedding` — a fixed-size vector of floats holding an embedding of `text`, computed with the `all-MiniLM-L6-v2` sentence-transformers model (embed each `text` exactly as written; use that same model for every row and for the query below). Populate it with exactly these 8 rows:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| id | text |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`|----|------|`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 1 | Espresso machine with a 15-bar pump for rich, crema-topped coffee |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 2 | Burr coffee grinder for fresh, evenly ground espresso beans |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 3 | Pour-over coffee kettle with a gooseneck spout for slow brewing |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 4 | Reusable coffee filter for drip brewers and pour-over cones |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 5 | Waterproof hiking boots for rocky mountain trails |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 6 | Lightweight hiking backpack with a hydration sleeve for day trails |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 7 | Trekking poles for steep mountain trail descents |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`| 8 | Merino wool hiking socks for long days on the trail |`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Then run two vector searches against the table, both using cosine distance: (1) a 'similar items' search using row 2's own stored embedding as the query, returning the 3 nearest OTHER rows; (2) a text search — embed the query `gear for a mountain hike` with the same model and return the 3 nearest rows. End your answer with two lines of exactly this form, ids comma-separated, nearest first:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`NEAREST: \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`QUERY_HITS: \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1002: build all indexes`,id:`1002-build-all-indexes`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The `eval_index_all` table is about to back our product search page. The page will run: visual similarity search on the `image_clip_v2` embeddings (cosine distance), keyword search over `body_text`, exact-match filters on `category`, and range filters on `price_usd` — often combined with the searches. Today the table has no indexes at all. Build all the indexes this table will need to make search fast, and make sure every one of them finishes building. Don't build indexes the page won't use — each one costs build time and slows writes — and note that `image_clip_v1` is a legacy column nothing queries anymore. Do not modify the table's rows or schema.\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1003: embedding models in branches`,id:`1003-embedding-models-in-branches`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"We're choosing an embedding model for the `body_text` column of the `eval_embed_branches` table (24 product descriptions) and want to compare 3 candidates side by side without touching the production data. The candidates: `all-MiniLM-L6-v2`, `bge-small-en-v1.5`, and `gte-small`. Run the experiment on branches: for each candidate, create a branch of the table off main and, on that branch only, add a fixed-size float vector column called `body_embedding` holding that model's embedding of every row's `body_text` (embed the text exactly as stored). Then run the same search on each branch, using cosine distance — the 3 products most similar to the query `a paperback novel to read on the beach`, embedded with that branch's model — so we can compare what each model returns. Main must be left exactly as it is: no new column, no new rows, no other writes. End your answer with one line per branch, each of exactly this form (the model named the way its provider does, e.g. `all-MiniLM-L6-v2`; the three ids comma-separated, most similar first):\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`BRANCH: \u003cbranch_name> | \u003cembedding_model> | \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1004: parquet to lance embed`,id:`1004-parquet-to-lance-embed`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"Convert this Parquet dataset to Lance and add a vector column from the description field. The dataset is `products.parquet` in your working directory. Load it into a new table called `eval_parquet_products` in this database, keeping every Parquet column and row exactly as they are, and add a column `description_embedding`: a fixed-size vector of floats holding an embedding of each row's `description`, computed with a text-embedding model of your choice (the same model for every row).\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1005: rag chatbot citations`,id:`1005-rag-chatbot-citations`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`7`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`7`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Build a RAG chatbot over this table with LangChain. Return citations.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:'The table is `eval_rag_docs` in this database: help-center articles with `doc_id`, `title`, and `body`. The chatbot gets no LLM API credentials, so generate its answers with a model that runs locally, or answer extractively from the retrieved passages. When the chatbot is working, ask it: \"How long do I have to return a damaged item for a refund?\" In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. Include the chatbot\\'s prose answer above the result lines for manual inspection. End your answer with two lines of exactly this form. ANSWER must contain only the number of days as decimal digits, with no units or other text:'})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`ANSWER: \u003cnumber of days>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`CITATIONS: \u003cdoc_id>,\u003cdoc_id>,... (the articles the chatbot cited)`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1006: fastapi similar endpoint`,id:`1006-fastapi-similar-endpoint`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`6`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Add a recommend-similar-products endpoint to this FastAPI service.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:\"The service is `app.py` in your working directory; it serves the `eval_rec_products` table in this database. Add an endpoint that returns the products most similar to a given product, using the `image_clip_v2` vectors with either cosine or L2 distance and excluding the query product itself. Keep the existing endpoint working. Then run the service and call your endpoint for product 5, asking for 3 results. In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. End your answer with a line of exactly this form, ids nearest first:\"})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`SIMILAR: \u003cid>,\u003cid>,\u003cid>`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})}),_jsx(Accordion,{title:`1007: agent tool past conversations`,id:`1007-agent-tool-past-conversations`,children:_jsx(CodeBlock,{filename:``,numberOfLines:`6`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes vitesse-light catppuccin-mocha`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#1e1e2e`,color:`#393a34`,\"--shiki-dark\":`#cdd6f4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Give my agent a tool that queries this table for relevant past conversations.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:'The table is `eval_conversations` in this database, one row per past support conversation (`conversation_id`, `user_id`, `summary`). The agent serves one signed-in user at a time. Build the tool, then call it the way the agent would for user `u-ana` asking \"my card was charged twice\". In your final answer, include the complete source of every file you created or modified to implement the app or tool, with filenames and fenced code blocks, including the code used to invoke it. Include the commands you actually ran and their observed output so the implementation can be manually inspected. Replace credential values with placeholders. Put this material before the final result lines, and emit each required result line only once. End your answer with a line of exactly this form, listing the conversation ids your tool returned:'})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`CONVERSATIONS: \u003cid>,\u003cid>,...`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})})})]})]})}function MDXContent(props={}){const{wrapper:MDXLayout}={..._provideComponents(),...props.components};return MDXLayout?_jsx(MDXLayout,{...props,children:_jsx(_createMdxContent,{...props})}):_createMdxContent(props)}return{default:MDXContent};function _missingMdxReference(id,component){throw new Error(`Expected `+(component?`component`:`object`)+\" `\"+id+\"` to be defined: you likely forgot to import, pass, or provide it.\")}","frontmatter":{},"scope":{"codeStyling":{"theme":{"light":"vitesse-light","dark":"catppuccin-mocha"}},"config":{},"pageMetadata":{"title":"Agentic skills","sidebarTitle":"Agentic Skills","description":"Give a coding agent a current, maintained reference for the LanceDB APIs.","icon":"robot","keywords":["agents","plugin","claude code","codex","cursor","copilot"],"href":"/enterprise/v1-13-0/basics/agentic-skills"}}},"mdxSourceWithNoJs":null},"mdxExtracts":{"tableOfContents":[{"title":"The plugin","slug":"plugin","depth":2,"children":[]},{"title":"Automated prompt catalog","slug":"automated-prompts","depth":2,"children":[{"title":"Tables and schema design","slug":"agent-evals-1","depth":3,"children":[]},{"title":"Data ingestion and updates","slug":"agent-evals-2","depth":3,"children":[]},{"title":"Queries and search","slug":"agent-evals-3","depth":3,"children":[]},{"title":"Index management","slug":"agent-evals-4","depth":3,"children":[]},{"title":"Version history and tags","slug":"agent-evals-5","depth":3,"children":[]},{"title":"Column metadata and lineage","slug":"agent-evals-6","depth":3,"children":[]},{"title":"Namespaces","slug":"agent-evals-7","depth":3,"children":[]},{"title":"Table branches","slug":"agent-evals-8","depth":3,"children":[]},{"title":"Background jobs","slug":"agent-evals-9","depth":3,"children":[]},{"title":"Multi-step workflows","slug":"agent-evals-10","depth":3,"children":[]}]}],"codeExamples":{}},"jsLinks":["/_mintlify/custom/lancedb-bcbb4faf/17e8f66511a11582/static/scripts/cookie-consent.js"],"subdomain":"lancedb-bcbb4faf","siteName":"LanceDB","title":"Agentic skills - LanceDB","navMaps":{"firstHrefInVersion":{"1.14.0":{"href":"/enterprise/v1-14-0/basics/agentic-skills","matchLevel":1},"1.13.0":{"href":"/enterprise/v1-13-0/basics/agentic-skills","matchLevel":0},"1.12.0":{"href":"/enterprise/v1-12-0/basics/agentic-skills","matchLevel":1},"0.41.0":{"href":"/oss/v0-41-0","matchLevel":4}},"firstHrefInLanguage":{}},"lastModified":"2026-11-20T17:15:46.505Z","originalFileLocation":"enterprise/v1-13-0/basics/agentic-skills.mdx"},"headHtml":"\u003cscript data-mint-head>(function(){var d=document.documentElement,p=null;if(!p){try{p=localStorage.getItem(\"isDarkMode\")}catch(e){}if(p==='true')p='dark';else if(p==='false')p='light';if(p!=='light'\u0026\u0026p!=='dark'\u0026\u0026p!=='system')p=\"system\"}var r=p==='system'?(matchMedia(\"(prefers-color-scheme: dark)\").matches?'dark':'light'):p;d.classList.remove(r==='dark'?'light':'dark');d.classList.add(r);d.setAttribute(\"data-theme-preference\",p)})();\u003c/script>\u003cmeta data-mint-head name=\"description\" content=\"Give a coding agent a current, maintained reference for the LanceDB APIs.\">\u003cmeta data-mint-head name=\"application-name\" content=\"LanceDB\">\u003cmeta data-mint-head name=\"generator\" content=\"Mintlify\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/android-chrome-192x192.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"apple-touch-icon\" type=\"image/png\" sizes=\"180x180\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/apple-touch-icon.png\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/favicon-16x16.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/favicon-32x32.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/favicon.ico\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon-dark/android-chrome-192x192.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon-dark/favicon-16x16.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon-dark/favicon-32x32.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon-dark/favicon.ico\" media=\"(prefers-color-scheme: dark)\">\u003cmeta data-mint-head property=\"og:title\" content=\"Agentic skills - LanceDB\">\u003cmeta data-mint-head property=\"og:description\" content=\"Give a coding agent a current, maintained reference for the LanceDB APIs.\">\u003cmeta data-mint-head property=\"og:site_name\" content=\"LanceDB\">\u003cmeta data-mint-head property=\"og:url\" content=\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills\">\u003cmeta data-mint-head property=\"og:type\" content=\"website\">\u003cmeta data-mint-head property=\"og:image\" content=\"https://lancedb-bcbb4faf.mintlify.site/_mintlify/api/og/v2?division=Getting+Started\u0026amp;title=Agentic+skills\u0026amp;description=Give+a+coding+agent+a+current%2C+maintained+reference+for+the+LanceDB+APIs.\u0026amp;theme=9d8045ca9469df4ceece463f\u0026amp;signature=b6af2fe27273ddf0cb0c8c0c469d4a4f099b3291dd8f34f3d4cd52bf0db0632c\">\u003cmeta data-mint-head property=\"og:image:width\" content=\"1200\">\u003cmeta data-mint-head property=\"og:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"twitter:card\" content=\"summary_large_image\">\u003cmeta data-mint-head name=\"twitter:title\" content=\"Agentic skills - LanceDB\">\u003cmeta data-mint-head name=\"twitter:description\" content=\"Give a coding agent a current, maintained reference for the LanceDB APIs.\">\u003cmeta data-mint-head name=\"twitter:image\" content=\"https://lancedb-bcbb4faf.mintlify.site/_mintlify/api/og/v2?division=Getting+Started\u0026amp;title=Agentic+skills\u0026amp;description=Give+a+coding+agent+a+current%2C+maintained+reference+for+the+LanceDB+APIs.\u0026amp;theme=9d8045ca9469df4ceece463f\u0026amp;signature=b6af2fe27273ddf0cb0c8c0c469d4a4f099b3291dd8f34f3d4cd52bf0db0632c\">\u003cmeta data-mint-head name=\"twitter:image:width\" content=\"1200\">\u003cmeta data-mint-head name=\"twitter:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"msapplication-config\" content=\"/mintlify-assets/_mintlify/favicons/lancedb-bcbb4faf/AJCI1CXWoYDPZbb4/_generated/favicon/browserconfig.xml\">\u003cmeta data-mint-head name=\"apple-mobile-web-app-title\" content=\"LanceDB\">\u003cmeta data-mint-head name=\"msapplication-TileColor\" content=\"#FF6B35\">\u003clink data-mint-head rel=\"canonical\" href=\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills\">\u003clink data-mint-head rel=\"alternate\" type=\"application/xml\" href=\"/sitemap.xml\">\u003clink data-mint-head rel=\"alternate\" type=\"text/markdown\" href=\"/enterprise/v1-13-0/basics/agentic-skills.md\">\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@graph\":[{\"@type\":\"Organization\",\"@id\":\"https://docs.lancedb.com/#organization\",\"name\":\"LanceDB\",\"url\":\"https://docs.lancedb.com\",\"logo\":{\"@type\":\"ImageObject\",\"url\":\"https://mintcdn.com/lancedb-bcbb4faf/6L0IRVkfdlgMU1Pw/static/assets/logo/dark-lancedb-logo.svg?fit=max\u0026auto=format\u0026n=6L0IRVkfdlgMU1Pw\u0026q=85\u0026s=d0592e59a29df2477080a857e13e341f\"}},{\"@type\":\"WebSite\",\"@id\":\"https://docs.lancedb.com#website\",\"name\":\"LanceDB\",\"url\":\"https://docs.lancedb.com\",\"publisher\":{\"@id\":\"https://docs.lancedb.com/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills#webpage\",\"url\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills\",\"name\":\"Agentic skills\",\"description\":\"Give a coding agent a current, maintained reference for the LanceDB APIs.\",\"dateModified\":\"2026-11-20T17:15:46.505Z\",\"isPartOf\":{\"@id\":\"https://docs.lancedb.com#website\"},\"breadcrumb\":{\"@id\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills#breadcrumb\"}},{\"@type\":\"BreadcrumbList\",\"@id\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Getting Started\",\"item\":\"https://docs.lancedb.com/enterprise/v1-13-0/quickstart\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Agentic skills\",\"item\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills\"}]},{\"@type\":[\"Article\",\"TechArticle\"],\"@id\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills#article\",\"headline\":\"Agentic skills\",\"name\":\"Agentic skills\",\"description\":\"Give a coding agent a current, maintained reference for the LanceDB APIs.\",\"url\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills\",\"mainEntityOfPage\":{\"@id\":\"https://docs.lancedb.com/enterprise/v1-13-0/basics/agentic-skills#webpage\"},\"dateModified\":\"2026-11-20T17:15:46.505Z\",\"keywords\":\"agents, plugin, claude code, codex, cursor, copilot\",\"publisher\":{\"@id\":\"https://docs.lancedb.com/#organization\"},\"isPartOf\":{\"@id\":\"https://docs.lancedb.com#website\"}}]}\u003c/script>\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@type\":\"WebSite\",\"name\":\"LanceDB\",\"publisher\":{\"@type\":\"Organization\",\"name\":\"LanceDB\"},\"provider\":{\"@type\":\"Organization\",\"name\":\"Mintlify\",\"url\":\"https://mintlify.com\"}}\u003c/script>","agentPointerHref":"/enterprise/v1-13-0/basics/agentic-skills.md","siteName":"LanceDB","nav":{"global":null,"products":[{"product":"Enterprise","icon":"server","description":"Multimodal lakehouse","versions":[{"version":"1.14.0","default":true,"tag":"Latest","pages":[{"title":"1.14.0","href":"/enterprise/v1-14-0/basics/agentic-skills"}]},{"version":"1.13.0","tabs":[{"tab":"Documentation","groups":[{"group":"Introduction","pages":[{"title":"LanceDB","sidebarTitle":"LanceDB","icon":"https://mintlify.s3.us-west-1.amazonaws.com/lancedb-bcbb4faf/enterprise/v1-13-0/static/assets/logo/lancedb-icon-gray.svg","href":"/enterprise/v1-13-0"},{"title":"LanceDB Embedded","sidebarTitle":"LanceDB Embedded","icon":"cube","href":"/enterprise/v1-13-0/basics/lancedb-embedded"},{"title":"LanceDB Enterprise","sidebarTitle":"LanceDB Enterprise","icon":"server","href":"/enterprise/v1-13-0/enterprise/index"},{"title":"Embedded vs Enterprise","sidebarTitle":"Embedded vs Enterprise","icon":"scale-balanced","href":"/enterprise/v1-13-0/basics/embedded-vs-enterprise"},{"title":"About this version","icon":"circle-info","href":"/enterprise/v1-13-0/about-this-version"}]},{"group":"Getting Started","pages":[{"title":"Quickstart","sidebarTitle":"Quickstart","icon":"rocket","href":"/enterprise/v1-13-0/quickstart"},{"title":"SDKs","sidebarTitle":"SDKs","icon":"code-compare","href":"/enterprise/v1-13-0/basics/sdk"},{"title":"SQL","sidebarTitle":"SQL","icon":"table-list","href":"/enterprise/v1-13-0/basics/sql"},{"title":"Console","sidebarTitle":"Console","icon":"window","href":"/enterprise/v1-13-0/enterprise/console"},{"title":"Import a Hugging Face dataset","description":"Store a Hugging Face token and import dataset rows into LanceDB Enterprise with SQL or Python.","icon":"file-import","href":"/enterprise/v1-13-0/enterprise/sql-hugging-face"},{"title":"Agentic skills","sidebarTitle":"Agentic Skills","icon":"robot","href":"/enterprise/v1-13-0/basics/agentic-skills"},{"title":"Support","sidebarTitle":"Support","description":"Where to get help: community Discord and email for LanceDB Embedded, dedicated support for LanceDB Enterprise.","icon":"circle-question","href":"/enterprise/v1-13-0/deployment/support"}]},{"group":"Concepts","pages":[{"title":"Lance format","sidebarTitle":"Lance Format","icon":"https://mintlify.s3.us-west-1.amazonaws.com/lancedb-bcbb4faf/enterprise/v1-13-0/static/assets/logo/lance-logo-gray.png","href":"/enterprise/v1-13-0/basics/lance"},{"title":"LanceDB Lakehouse","sidebarTitle":"LanceDB Lakehouse","icon":"sitemap","href":"/enterprise/v1-13-0/basics/catalog"},{"title":"Job Framework","sidebarTitle":"Job Framework","icon":"hourglass-half","href":"/enterprise/v1-13-0/basics/jobs"},{"title":"Realtime Engine (RTE)","sidebarTitle":"Realtime Engine (RTE)","icon":"bolt","href":"/enterprise/v1-13-0/tables/realtime-engine"},{"title":"Consistency","sidebarTitle":"Consistency","icon":"water","href":"/enterprise/v1-13-0/tables/consistency"},{"title":"Guarantees","sidebarTitle":"Guarantees","icon":"shield-check","href":"/enterprise/v1-13-0/basics/guarantees"},{"title":"Limits","sidebarTitle":"Limits","icon":"ruler","href":"/enterprise/v1-13-0/basics/limits"},{"title":"Naming rules","sidebarTitle":"Naming Rules","icon":"signature","href":"/enterprise/v1-13-0/basics/naming"}]},{"group":"Data Ingestion","pages":[{"title":"Ingesting data","sidebarTitle":"Overview","icon":"inbox","href":"/enterprise/v1-13-0/ingestion/index"},{"title":"Creating a table","sidebarTitle":"Create Table","icon":"table","href":"/enterprise/v1-13-0/ingestion/create-table"},{"title":"Insert data","sidebarTitle":"Insert","icon":"plus","href":"/enterprise/v1-13-0/ingestion/insert"},{"title":"Update rows","sidebarTitle":"Update","icon":"pen","href":"/enterprise/v1-13-0/ingestion/update"},{"title":"Delete rows","sidebarTitle":"Delete","icon":"trash","href":"/enterprise/v1-13-0/ingestion/delete"},{"title":"RTE Write (Upsert)","sidebarTitle":"RTE Write (Upsert)","icon":"bolt","href":"/enterprise/v1-13-0/ingestion/realtime-write"},{"title":"Define Storage Integration","sidebarTitle":"Define Storage Integration","icon":"cloud","href":"/enterprise/v1-13-0/ingestion/define-storage-integration"},{"title":"Attach Database","sidebarTitle":"Attach Database","icon":"plug","href":"/enterprise/v1-13-0/ingestion/attach-database"},{"title":"Import Datasets","sidebarTitle":"Import Datasets","icon":"file-import","href":"/enterprise/v1-13-0/ingestion/import"},{"title":"Distributed Write","sidebarTitle":"Distributed Write","icon":"upload","href":"/enterprise/v1-13-0/enterprise/distributed-write"},{"title":"Distributed RTE Write","sidebarTitle":"Distributed RTE Write","icon":"bolt","href":"/enterprise/v1-13-0/enterprise/distributed-real-time-write"}]},{"group":"Search","pages":[{"title":"Search","sidebarTitle":"Overview","icon":"list","href":"/enterprise/v1-13-0/search/index"},{"title":"Vector Search","sidebarTitle":"Vector Search","icon":"arrow-up-right-dots","href":"/enterprise/v1-13-0/search/vector-search"},{"title":"Full-Text Search (FTS)","sidebarTitle":"Full-Text Search (FTS)","icon":"book","href":"/enterprise/v1-13-0/search/full-text-search"},{"title":"Hybrid Search","sidebarTitle":"Hybrid Search","icon":"search","href":"/enterprise/v1-13-0/search/hybrid-search"},{"title":"Metadata Filtering in LanceDB","sidebarTitle":"Filtering","icon":"filter","href":"/enterprise/v1-13-0/search/filtering"},{"title":"Reranking Search Results","sidebarTitle":"Reranking","icon":"sort-amount-down","href":"/enterprise/v1-13-0/reranking/index"},{"title":"RTE reads","sidebarTitle":"RTE Read","icon":"magnifying-glass","href":"/enterprise/v1-13-0/search/realtime-read"},{"title":"Distributed RTE Read","sidebarTitle":"Distributed RTE Read","icon":"network-wired","href":"/enterprise/v1-13-0/enterprise/distributed-real-time-read"},{"title":"Distributed Search","sidebarTitle":"Distributed Search","icon":"network-wired","href":"/enterprise/v1-13-0/enterprise/distributed-search"},{"title":"Data Page Cache","sidebarTitle":"Data Page Cache","icon":"database","href":"/enterprise/v1-13-0/enterprise/data-page-cache"},{"title":"Tiered Index Cache","sidebarTitle":"Tiered Index Cache","icon":"layer-group","href":"/enterprise/v1-13-0/enterprise/tiered-index-cache"},{"title":"Cache Prewarm","sidebarTitle":"Cache Prewarm","icon":"fire","href":"/enterprise/v1-13-0/enterprise/cache-prewarm"}]},{"group":"Data Curation","pages":[{"title":"Data curation","sidebarTitle":"Overview","icon":"filter","href":"/enterprise/v1-13-0/curation/index"},{"title":"Inspect Datasets","sidebarTitle":"Inspect Datasets","icon":"database","href":"/enterprise/v1-13-0/curation/inspect-datasets"},{"title":"Create a Function","tag":"Early access","sidebarTitle":"Create Function","icon":"code","href":"/enterprise/v1-13-0/curation/create-function"},{"title":"Choose how to compute a column","sidebarTitle":"Function authoring guidance","icon":"file-code","tag":"Early access","href":"/enterprise/v1-13-0/enterprise/function-authoring"},{"title":"Operate Function workloads","sidebarTitle":"Function Operations","icon":"gears","tag":"Early access","href":"/enterprise/v1-13-0/enterprise/function-operations"},{"title":"Function worker availability","sidebarTitle":"Function Workers","icon":"server","tag":"Early access","href":"/enterprise/v1-13-0/enterprise/deployment/function-workers"},{"title":"Chunk records into multiple rows","sidebarTitle":"Chunking","icon":"diagram-subtask","href":"/enterprise/v1-13-0/curation/chunking"},{"title":"Statistical Sampling","sidebarTitle":"Statistical Sampling","icon":"dice","href":"/enterprise/v1-13-0/curation/statistical-sampling"},{"title":"Selective Sampling","sidebarTitle":"Selective Sampling","icon":"crosshairs","href":"/enterprise/v1-13-0/curation/selective-sampling"},{"title":"Profiling a dataset","sidebarTitle":"Aggregate","icon":"chart-column","href":"/enterprise/v1-13-0/curation/aggregate"},{"title":"Joining datasets","sidebarTitle":"Join","icon":"link","href":"/enterprise/v1-13-0/curation/join"},{"title":"Blob Late Materialization","sidebarTitle":"Blob Late Materialization","icon":"clock","href":"/enterprise/v1-13-0/curation/blob-late-materialization"},{"title":"Distributed Shuffle","sidebarTitle":"Distributed Shuffle","icon":"shuffle","href":"/enterprise/v1-13-0/curation/distributed-shuffle"},{"title":"Ranking","sidebarTitle":"Ranking","icon":"chart-simple","href":"/enterprise/v1-13-0/curation/windowing"},{"title":"Cluster records by index partition","sidebarTitle":"Clustering","icon":"circle-nodes","href":"/enterprise/v1-13-0/curation/clustering"},{"title":"Deduplicate near-duplicate records","sidebarTitle":"Dedupe","icon":"clone","href":"/enterprise/v1-13-0/curation/dedupe"},{"title":"Scope curation to what changed","sidebarTitle":"Changes","icon":"clock-rotate-left","href":"/enterprise/v1-13-0/curation/changes"},{"title":"Create a view","sidebarTitle":"Create View","icon":"eye","href":"/enterprise/v1-13-0/feature-engineering/view"},{"title":"Vend a Secret to a Function","tag":"Early access","sidebarTitle":"Vend Secret","icon":"key","href":"/enterprise/v1-13-0/curation/vend-secret"},{"title":"Download S3 objects with a Function","icon":"cloud-arrow-down","tag":"Early access","href":"/enterprise/v1-13-0/enterprise/function-s3-download"},{"title":"Use Secrets with Python Functions","sidebarTitle":"Secrets walkthrough","icon":"key","tag":"Early access","href":"/enterprise/v1-13-0/enterprise/function-secrets"}]},{"group":"Feature Engineering","pages":[{"title":"Feature engineering","sidebarTitle":"Overview","icon":"layer-group","href":"/enterprise/v1-13-0/feature-engineering/index"},{"title":"Create a computed column","sidebarTitle":"Create Computed Column","icon":"table-columns","href":"/enterprise/v1-13-0/feature-engineering/add-computed-column"},{"title":"Refresh a computed column","sidebarTitle":"Refresh Computed Column","icon":"arrows-rotate","href":"/enterprise/v1-13-0/feature-engineering/refresh-column"},{"title":"Create a materialized view","sidebarTitle":"Create Materialized View","icon":"table","href":"/enterprise/v1-13-0/feature-engineering/materialized-views"},{"title":"Refresh a materialized view","sidebarTitle":"Refresh Materialized View","icon":"arrows-rotate","href":"/enterprise/v1-13-0/feature-engineering/refresh-materialized-view"},{"title":"Incremental Refresh","sidebarTitle":"Incremental Refresh","icon":"arrow-rotate-right","href":"/enterprise/v1-13-0/enterprise/incremental-refresh"},{"title":"Auto Refresh","sidebarTitle":"Auto Refresh","icon":"clock-rotate-left","href":"/enterprise/v1-13-0/enterprise/auto-refresh"},{"title":"Function Refresh Checkpointing","tag":"Early access","sidebarTitle":"Checkpointing","icon":"floppy-disk","href":"/enterprise/v1-13-0/enterprise/checkpointing"},{"title":"Adaptive Execution","tag":"Early access","sidebarTitle":"Adaptive Execution","icon":"gauge-high","href":"/enterprise/v1-13-0/enterprise/adaptive-execution"},{"group":"Geneva","icon":"layer-group","pages":[{"title":"Multimodal Feature Engineering with Geneva","sidebarTitle":"Overview","icon":"layer-group","href":"/enterprise/v1-13-0/geneva/index"},{"title":"What is Feature Engineering?","sidebarTitle":"What is feature engineering?","icon":"grip-horizontal","href":"/enterprise/v1-13-0/geneva/overview/index"},{"title":"Getting Started","sidebarTitle":"Getting Started","icon":"rocket","href":"/enterprise/v1-13-0/geneva/getting-started"},{"group":"Transforms","pages":[{"title":"Understanding Transforms","sidebarTitle":"Understanding Transforms","icon":"code-compare","href":"/enterprise/v1-13-0/geneva/udfs/index"},{"title":"User-Defined Functions (UDFs)","sidebarTitle":"UDFs","icon":"function","href":"/enterprise/v1-13-0/geneva/udfs/udfs"},{"title":"Chunkers (Scalar UDTFs)","sidebarTitle":"Chunkers","icon":"diagram-subtask","href":"/enterprise/v1-13-0/geneva/udfs/scalar-udtfs"},{"title":"Batch User-Defined Table Functions (UDTFs)","sidebarTitle":"Batch UDTFs","icon":"layer-group","href":"/enterprise/v1-13-0/geneva/udfs/batch-udtfs"},{"title":"Error Handling in Geneva UDFs","sidebarTitle":"Error Handling","icon":"exclamation","href":"/enterprise/v1-13-0/geneva/udfs/error_handling"},{"title":"Profiling Stateful UDF Memory","sidebarTitle":"Profiling Memory","icon":"chart-line","href":"/enterprise/v1-13-0/geneva/udfs/profiling-memory"},{"title":"Blob Types in Geneva UDFs","sidebarTitle":"Blobs","icon":"splotch","href":"/enterprise/v1-13-0/geneva/udfs/blobs"}]},{"group":"Built-in Transforms","pages":[{"title":"Built-in LLM and Embedding UDFs","sidebarTitle":"Overview","icon":"sparkles","href":"/enterprise/v1-13-0/geneva/udfs/providers/index"},{"title":"OpenAI","sidebarTitle":"OpenAI","icon":"brain","href":"/enterprise/v1-13-0/geneva/udfs/providers/openai"},{"title":"Gemini","sidebarTitle":"Gemini","icon":"google","href":"/enterprise/v1-13-0/geneva/udfs/providers/gemini"},{"title":"Sentence Transformers","sidebarTitle":"Sentence Transformers","icon":"microchip","href":"/enterprise/v1-13-0/geneva/udfs/providers/sentence-transformers"}]},{"group":"Running Jobs","pages":[{"title":"Geneva Jobs","sidebarTitle":"Job execution","href":"/enterprise/v1-13-0/geneva/jobs/index"},{"title":"Backfilling","sidebarTitle":"Backfilling","icon":"fill","href":"/enterprise/v1-13-0/geneva/jobs/backfilling"},{"title":"Bulk Loading \u0026 Updating Columns","sidebarTitle":"Bulk Load / Update Columns","icon":"columns-3","href":"/enterprise/v1-13-0/geneva/jobs/bulk-load-columns"},{"title":"Materialized Views with UDFs","sidebarTitle":"Materialized views","icon":"microscope","href":"/enterprise/v1-13-0/geneva/jobs/materialized-views"},{"title":"Advanced Job Configuration","sidebarTitle":"Advanced job configuration","icon":"sliders","href":"/enterprise/v1-13-0/geneva/jobs/advanced-job-configuration"},{"title":"Job Lifecycle","sidebarTitle":"Job Lifecycle","icon":"arrows-spin","href":"/enterprise/v1-13-0/geneva/jobs/lifecycle"},{"title":"Backfill Conflicts","sidebarTitle":"Backfill Conflicts","icon":"code-merge","href":"/enterprise/v1-13-0/geneva/jobs/conflicts"},{"title":"Distributed Job Performance and Cluster Sizing","sidebarTitle":"Performance and Sizing","icon":"chart-line","href":"/enterprise/v1-13-0/geneva/jobs/performance"},{"title":"Job Metrics (Diagnostics)","sidebarTitle":"Job Metrics","icon":"chart-simple","href":"/enterprise/v1-13-0/geneva/jobs/job_metrics"},{"title":"Geneva Console","sidebarTitle":"Console","icon":"computer","href":"/enterprise/v1-13-0/geneva/jobs/console"},{"title":"Troubleshooting Geneva Jobs","sidebarTitle":"Job troubleshooting","icon":"wrench","href":"/enterprise/v1-13-0/geneva/jobs/troubleshooting"},{"title":"Advanced Execution Contexts","sidebarTitle":"Advanced execution contexts","icon":"circle-nodes","href":"/enterprise/v1-13-0/geneva/jobs/contexts"}]},{"group":"Deployment","pages":[{"title":"Deploy Geneva using Helm","sidebarTitle":"Helm Deployment","icon":"cogs","href":"/enterprise/v1-13-0/geneva/deployment/helm"},{"title":"Job and Session Startup Optimizations","sidebarTitle":"Startup optimizations","icon":"engine","href":"/enterprise/v1-13-0/geneva/jobs/startup"},{"title":"Dependency Verification","sidebarTitle":"Dependency Verification","icon":"magnifying-glass-chart","href":"/enterprise/v1-13-0/geneva/deployment/dependency-verification"},{"title":"Advanced Configuration","sidebarTitle":"Advanced Configuration","icon":"flask-gear","href":"/enterprise/v1-13-0/geneva/udfs/advanced-configuration"},{"title":"Manual Deployment on Kubernetes","sidebarTitle":"Manual Deployment","icon":"cogs","href":"/enterprise/v1-13-0/geneva/deployment/index"},{"title":"Troubleshooting Geneva Deployments","sidebarTitle":"Deployment Troubleshooting","icon":"question-circle","href":"/enterprise/v1-13-0/geneva/deployment/troubleshooting"}]},{"title":"End-to-End Example","sidebarTitle":"End-to-end example","icon":"play","href":"/enterprise/v1-13-0/geneva/end-to-end"},{"title":"Geneva Python SDK","icon":"python","href":"/enterprise/v1-13-0/geneva/reference"}]}]},{"group":"Training","pages":[{"title":"PyTorch Data Loader","sidebarTitle":"PyTorch Data Loader","icon":"boxes-stacked","href":"/enterprise/v1-13-0/training/data-loading"},{"title":"Training Cache","sidebarTitle":"Training Cache","icon":"gauge-high","href":"/enterprise/v1-13-0/training/training-cache"},{"title":"Permutation","sidebarTitle":"Permutation","icon":"shuffle","href":"/enterprise/v1-13-0/training/permutation"},{"title":"Observability","sidebarTitle":"Observability","icon":"gauge","href":"/enterprise/v1-13-0/training/observability"},{"title":"Distributed Training","sidebarTitle":"Distributed Training","icon":"server","href":"/enterprise/v1-13-0/training/distributed-training"},{"title":"Checkpointing","sidebarTitle":"Checkpointing","icon":"floppy-disk","href":"/enterprise/v1-13-0/training/checkpointing"},{"title":"Resuming","sidebarTitle":"Resuming","icon":"rotate-left","href":"/enterprise/v1-13-0/training/resuming"}]},{"group":"Management","pages":[{"title":"Management","sidebarTitle":"Overview","icon":"folder-tree","href":"/enterprise/v1-13-0/management/index"},{"title":"Catalog","sidebarTitle":"Catalog","icon":"sitemap","href":"/enterprise/v1-13-0/management/catalog"},{"title":"Database","sidebarTitle":"Database","icon":"database","href":"/enterprise/v1-13-0/management/database"},{"title":"Namespaces","sidebarTitle":"Namespace","icon":"folder-tree","href":"/enterprise/v1-13-0/management/namespace"},{"title":"Tables","sidebarTitle":"Table","icon":"table","href":"/enterprise/v1-13-0/management/table"},{"title":"Compress Function output columns","icon":"compress","href":"/enterprise/v1-13-0/enterprise/function-column-compression"},{"title":"Views","sidebarTitle":"View","icon":"eye","href":"/enterprise/v1-13-0/management/view"},{"title":"Materialized view","sidebarTitle":"Materialized View","icon":"table","href":"/enterprise/v1-13-0/management/materialized-view"},{"title":"Secret","tag":"Early access","sidebarTitle":"Secret","icon":"key","href":"/enterprise/v1-13-0/management/secret"},{"title":"Functions","tag":"Early access","sidebarTitle":"Function","icon":"code","href":"/enterprise/v1-13-0/management/function"},{"title":"Handle Function failures","sidebarTitle":"Function Failures","icon":"triangle-exclamation","tag":"Early access","href":"/enterprise/v1-13-0/management/function-failures"},{"title":"Storage Integration","sidebarTitle":"Storage Integration","icon":"cloud","href":"/enterprise/v1-13-0/management/storage-integration"}]},{"group":"Maintenance","pages":[{"title":"Indexing Data","sidebarTitle":"Overview","icon":"list","href":"/enterprise/v1-13-0/indexing/index"},{"title":"Intelligent Auto-Maintenance","sidebarTitle":"Intelligent Auto-Maintenance","icon":"brain","href":"/enterprise/v1-13-0/enterprise/intelligent-auto-maintenance"},{"title":"Vector Indexing","sidebarTitle":"Vector Indexing","icon":"arrow-up-right-dots","href":"/enterprise/v1-13-0/indexing/vector-index"},{"title":"Full-Text Search (FTS) Indexing","sidebarTitle":"FTS Indexing","icon":"book","href":"/enterprise/v1-13-0/indexing/fts-index"},{"title":"Scalar Indexing","sidebarTitle":"Scalar Indexing","icon":"tree","href":"/enterprise/v1-13-0/indexing/scalar-index"},{"title":"Incremental Indexing","sidebarTitle":"Incremental Indexing","icon":"refresh","href":"/enterprise/v1-13-0/indexing/incremental-indexing"},{"title":"Distributed Indexing","sidebarTitle":"Distributed Indexing","icon":"layer-group","href":"/enterprise/v1-13-0/enterprise/distributed-indexing"},{"title":"GPU Indexing","sidebarTitle":"GPU Indexing","icon":"microchip","href":"/enterprise/v1-13-0/indexing/gpu-indexing"},{"title":"Compaction and cleanup","sidebarTitle":"Compaction","icon":"broom","href":"/enterprise/v1-13-0/maintenance/compaction"},{"title":"Realtime Engine compaction","sidebarTitle":"RTE Compaction","icon":"compress","href":"/enterprise/v1-13-0/enterprise/realtime-engine-compaction"},{"title":"Distributed Compaction","sidebarTitle":"Distributed Compaction","icon":"compress","href":"/enterprise/v1-13-0/enterprise/distributed-compaction"},{"title":"Advanced Conflict Resolution","sidebarTitle":"Advanced Conflict Resolution","icon":"shuffle","href":"/enterprise/v1-13-0/enterprise/advanced-conflict-resolution"},{"title":"Garbage Collection","sidebarTitle":"Garbage Collection","icon":"broom","href":"/enterprise/v1-13-0/maintenance/cleanup"}]},{"group":"Security","pages":[{"title":"Authentication","sidebarTitle":"Authentication","icon":"key","href":"/enterprise/v1-13-0/enterprise/authentication"},{"title":"Authorization","sidebarTitle":"Authorization","icon":"shield-halved","href":"/enterprise/v1-13-0/governance/authorization"},{"title":"Principals","sidebarTitle":"Principals","icon":"user","href":"/enterprise/v1-13-0/governance/principals"},{"title":"Privileges","sidebarTitle":"Privileges","icon":"key","href":"/enterprise/v1-13-0/governance/privileges"},{"title":"Audit Log","sidebarTitle":"Audit Log","icon":"clipboard-list","href":"/enterprise/v1-13-0/governance/audit-log"},{"title":"Compliance","sidebarTitle":"Compliance","icon":"shield-alt","href":"/enterprise/v1-13-0/enterprise/security"}]},{"group":"Deployment","pages":[{"title":"Horizontal Scaling","sidebarTitle":"Horizontal Scaling","icon":"arrows-left-right","href":"/enterprise/v1-13-0/enterprise/horizontal-scaling"},{"title":"Monitor LanceDB with OpenTelemetry","sidebarTitle":"OpenTelemetry","icon":"chart-line","href":"/enterprise/v1-13-0/deployment/opentelemetry"},{"title":"Kubernetes deployment guide","sidebarTitle":"Kubernetes","icon":"cloud","href":"/enterprise/v1-13-0/enterprise/deployment/kubernetes"},{"title":"Multicloud","sidebarTitle":"Multicloud","icon":"cloud","href":"/enterprise/v1-13-0/deployment/multicloud"},{"title":"Slurm","sidebarTitle":"Slurm","icon":"server","href":"/enterprise/v1-13-0/deployment/slurm"}]},{"group":"SQL Reference","pages":[{"title":"SQL reference","sidebarTitle":"Overview","icon":"table-list","href":"/enterprise/v1-13-0/sql/index"},{"title":"SQL identifiers and strings","sidebarTitle":"Identifier quoting","icon":"code","href":"/enterprise/v1-13-0/sql/identifiers"},{"title":"Catalog listing pagination","icon":"list-ol","href":"/enterprise/v1-13-0/sql/pagination"},{"title":"SQL data types","sidebarTitle":"Data Types","icon":"shapes","href":"/enterprise/v1-13-0/sql/data-types"},{"group":"DDL","icon":"sitemap","pages":[{"title":"Database","sidebarTitle":"Database","icon":"database","href":"/enterprise/v1-13-0/sql/ddl/database"},{"title":"Namespace","sidebarTitle":"Namespace","icon":"folder-tree","href":"/enterprise/v1-13-0/sql/ddl/namespace"},{"title":"Table","sidebarTitle":"Table","icon":"table","href":"/enterprise/v1-13-0/sql/ddl/table"},{"title":"View","sidebarTitle":"View","icon":"eye","href":"/enterprise/v1-13-0/sql/ddl/view"},{"title":"Materialized View","sidebarTitle":"Materialized View","icon":"table","href":"/enterprise/v1-13-0/sql/ddl/materialized-view"},{"title":"Secret","tag":"Early access","sidebarTitle":"Secret","icon":"key","href":"/enterprise/v1-13-0/sql/ddl/secret"},{"title":"Function","tag":"Early access","sidebarTitle":"Function","icon":"function","href":"/enterprise/v1-13-0/sql/ddl/function"},{"title":"Storage Integration","sidebarTitle":"Storage Integration","icon":"cloud","href":"/enterprise/v1-13-0/sql/ddl/storage-integration"},{"title":"Index","sidebarTitle":"Index","icon":"bolt","href":"/enterprise/v1-13-0/sql/ddl/index"},{"title":"Job","sidebarTitle":"Job","icon":"hourglass-half","href":"/enterprise/v1-13-0/sql/ddl/job"},{"title":"Grant","sidebarTitle":"Grant","icon":"user-lock","href":"/enterprise/v1-13-0/sql/ddl/grant"}]},{"group":"DML","icon":"database","pages":[{"title":"Insert","sidebarTitle":"Insert","icon":"plus","href":"/enterprise/v1-13-0/sql/dml/insert"},{"title":"Update","sidebarTitle":"Update","icon":"pen","href":"/enterprise/v1-13-0/sql/dml/update"},{"title":"Delete","sidebarTitle":"Delete","icon":"trash","href":"/enterprise/v1-13-0/sql/dml/delete"},{"title":"Copy","sidebarTitle":"Copy","icon":"file-export","href":"/enterprise/v1-13-0/sql/dml/copy"},{"title":"REFRESH COLUMN","sidebarTitle":"REFRESH COLUMN","icon":"arrows-rotate","href":"/enterprise/v1-13-0/sql/dml/refresh-column"},{"title":"REFRESH MATERIALIZED VIEW","sidebarTitle":"REFRESH MATERIALIZED VIEW","icon":"arrows-rotate","href":"/enterprise/v1-13-0/sql/dml/refresh-materialized-view"},{"title":"REFRESH","sidebarTitle":"REFRESH","icon":"arrows-rotate","href":"/enterprise/v1-13-0/sql/dml/refresh"}]},{"group":"DQL","icon":"filter","pages":[{"title":"SELECT","sidebarTitle":"SELECT","icon":"table-columns","href":"/enterprise/v1-13-0/sql/dql/select"},{"title":"Subqueries","sidebarTitle":"Subqueries","icon":"code-branch","href":"/enterprise/v1-13-0/sql/dql/subqueries"},{"title":"Joins","sidebarTitle":"Joins","icon":"link","href":"/enterprise/v1-13-0/sql/dql/join"},{"title":"Operators and literals","sidebarTitle":"Operators","icon":"plus-minus","href":"/enterprise/v1-13-0/sql/dql/operators"},{"title":"Functions","sidebarTitle":"Functions","icon":"function","href":"/enterprise/v1-13-0/sql/dql/functions"},{"title":"Enterprise SQL function index","sidebarTitle":"Enterprise functions","icon":"square-root-variable","href":"/enterprise/v1-13-0/sql-reference/functions-all"},{"title":"Scalar functions","sidebarTitle":"Scalar Functions","icon":"calculator","href":"/enterprise/v1-13-0/sql/dql/scalar-functions"},{"title":"Aggregation","sidebarTitle":"Aggregation","icon":"layer-group","href":"/enterprise/v1-13-0/sql/dql/aggregation"},{"title":"Window functions","sidebarTitle":"Window Functions","icon":"chart-simple","href":"/enterprise/v1-13-0/sql/dql/window"},{"title":"Random sampling","sidebarTitle":"Sampling","icon":"dice","href":"/enterprise/v1-13-0/sql/dql/sampling"},{"title":"Changes since a version","sidebarTitle":"Changes","icon":"clock-rotate-left","href":"/enterprise/v1-13-0/sql/dql/changes"},{"title":"Full-Text Search","sidebarTitle":"Full-Text Search","icon":"magnifying-glass","href":"/enterprise/v1-13-0/sql/dql/fts"},{"title":"CONTAINS_TOKENS","sidebarTitle":"CONTAINS_TOKENS","icon":"magnifying-glass","href":"/enterprise/v1-13-0/sql/dql/contains-tokens"},{"title":"IVF_PARTITION","sidebarTitle":"IVF_PARTITION","icon":"circle-nodes","href":"/enterprise/v1-13-0/sql/dql/ivf-partition"},{"title":"DEDUPE","sidebarTitle":"DEDUPE","icon":"layer-group","href":"/enterprise/v1-13-0/sql/dql/dedupe"},{"title":"DUPLICATE_PAIRS","sidebarTitle":"DUPLICATE_PAIRS","icon":"clone","href":"/enterprise/v1-13-0/sql/dql/duplicate-pairs"},{"title":"MinHash SQL functions","href":"/enterprise/v1-13-0/enterprise/sql-minhash"},{"title":"Vector distance functions","sidebarTitle":"Vector Distance","icon":"ruler","href":"/enterprise/v1-13-0/sql/dql/vector-distance"},{"title":"BLOB_DESCRIPTOR","sidebarTitle":"BLOB_DESCRIPTOR","icon":"file-binary","href":"/enterprise/v1-13-0/sql/dql/blob-descriptor"},{"title":"GENERATE_SERIES","sidebarTitle":"GENERATE_SERIES","icon":"list-ol","href":"/enterprise/v1-13-0/sql/dql/generate-series"},{"title":"Explain","sidebarTitle":"Explain","icon":"map","href":"/enterprise/v1-13-0/sql/dql/explain"}]}]}]},{"tab":"Integrations","pages":[{"title":"Integrations","href":"/enterprise/v1-13-0/integrations"}]},{"tab":"Tutorials","pages":[{"title":"Tutorials","href":"/enterprise/v1-13-0/tutorials"}]},{"tab":"Demos","pages":[{"title":"Demos","href":"/enterprise/v1-13-0/demos"}]},{"tab":"Datasets","pages":[{"title":"Datasets","href":"/enterprise/v1-13-0/datasets"}]},{"tab":"API Reference","pages":[{"title":"API Reference","href":"/enterprise/v1-13-0/api-reference/rest"}]}]},{"version":"1.12.0","pages":[{"title":"1.12.0","href":"/enterprise/v1-12-0/basics/agentic-skills"}]}]},{"product":"Embedded","icon":"database","description":"Open-source library","pages":[{"title":"Embedded","href":"/oss/v0-41-0"}]}]},"basePath":"","title":"Agentic skills - LanceDB","deploymentId":"dpl_vSEmv961Ncyvz1psVwgUnTXQMSYk","buildId":"dpl_vSEmv961Ncyvz1psVwgUnTXQMSYk","dynamics":["10e4515597"],"chunks":["BannerMarkdown-DnSAhvaM.js","dist-DIfYWxfm.js","lib-DMeTj1gb.js","lib-B8HbPgcC.js","katex-Dbtb6Wk7.js","html-void-elements-mCSPHezU.js","marked.esm-CZWythHk.js"]}