feat: provision + export documents table with generated SQL, SDK, and CLI

.agents/skills/cli-default/SKILL.md

@@ -1,13 +1,13 @@
 ---
 name: cli-default
-description: CLI tool (agentic-db) for the default API — provides CRUD commands for 91 tables and 3 custom operations
+description: CLI tool (agentic-db) for the default API — provides CRUD commands for 95 tables and 3 custom operations
 ---
 
 # cli-default
 
 <!-- @constructive-io/graphql-codegen - DO NOT EDIT -->
 
-CLI tool (agentic-db) for the default API — provides CRUD commands for 91 tables and 3 custom operations
+CLI tool (agentic-db) for the default API — provides CRUD commands for 95 tables and 3 custom operations
 
 ## Usage
 
@@ -76,6 +76,8 @@ See the `references/` directory for detailed per-entity API documentation:
 - [task](references/task.md)
 - [company](references/company.md)
 - [deal](references/deal.md)
+- [company-document](references/company-document.md)
+- [document](references/document.md)
 - [company-event](references/company-event.md)
 - [event](references/event.md)
 - [company-image](references/company-image.md)
@@ -103,6 +105,7 @@ See the `references/` directory for detailed per-entity API documentation:
 - [deal-company](references/deal-company.md)
 - [deal-contact](references/deal-contact.md)
 - [deal-note](references/deal-note.md)
+- [documents-chunk](references/documents-chunk.md)
 - [email-attachment](references/email-attachment.md)
 - [email-note](references/email-note.md)
 - [email-recipient](references/email-recipient.md)
@@ -121,6 +124,7 @@ See the `references/` directory for detailed per-entity API documentation:
 - [notes-chunk](references/notes-chunk.md)
 - [place](references/place.md)
 - [project-contact](references/project-contact.md)
+- [project-document](references/project-document.md)
 - [provider-sync-state](references/provider-sync-state.md)
 - [raw-contact](references/raw-contact.md)
 - [raw-contact-email](references/raw-contact-email.md)

.agents/skills/cli-default/references/company-document.md

@@ -0,0 +1,68 @@
+# companyDocument
+
+<!-- @constructive-io/graphql-codegen - DO NOT EDIT -->
+
+CRUD operations for CompanyDocument records via agentic-db CLI
+
+## Usage
+
+```bash
+agentic-db company-document list
+agentic-db company-document list --where.<field>.<op> <value> --orderBy <values>
+agentic-db company-document list --limit 10 --after <cursor>
+agentic-db company-document find-first --where.<field>.<op> <value>
+agentic-db company-document get --id <UUID>
+agentic-db company-document create --companyId <UUID> --documentId <UUID>
+agentic-db company-document update --id <UUID> [--companyId <UUID>] [--documentId <UUID>]
+agentic-db company-document delete --id <UUID>
+```
+
+## Examples
+
+### List companyDocument records
+
+```bash
+agentic-db company-document list
+```
+
+### List companyDocument records with pagination
+
+```bash
+agentic-db company-document list --limit 10 --offset 0
+```
+
+### List companyDocument records with cursor pagination
+
+```bash
+agentic-db company-document list --limit 10 --after <cursor>
+```
+
+### Find first matching companyDocument
+
+```bash
+agentic-db company-document find-first --where.id.equalTo <value>
+```
+
+### List companyDocument records with field selection
+
+```bash
+agentic-db company-document list --select id,id
+```
+
+### List companyDocument records with filtering and ordering
+
+```bash
+agentic-db company-document list --where.id.equalTo <value> --orderBy ID_ASC
+```
+
+### Create a companyDocument
+
+```bash
+agentic-db company-document create --companyId <UUID> --documentId <UUID>
+```
+
+### Get a companyDocument by id
+
+```bash
+agentic-db company-document get --id <value>
+```

.agents/skills/cli-default/references/document.md

@@ -4,31 +4,146 @@
 
 CRUD operations for Document records via agentic-db CLI
 
-**Unified Search API fields:** `contentBm25Score`, `embeddingTextBm25Score`, `titleTrgmSimilarity`, `urlTrgmSimilarity`, `contentTrgmSimilarity`, `sourceTypeTrgmSimilarity`, `abstractTrgmSimilarity`, `overviewTrgmSimilarity`, `embeddingTextTrgmSimilarity`, `searchScore`
+**pgvector embedding fields:** `embedding`
+High-dimensional vector columns for semantic similarity search. Query via the Unified Search API pgvector adapter using cosine, L2, or inner-product distance. Supports chunk-aware search: set `includeChunks: true` in VectorNearbyInput to transparently query across parent and chunk embeddings, returning the minimum distance.
+
+**Unified Search API fields:** `embeddingTextBm25Score`, `titleTrgmSimilarity`, `contentTrgmSimilarity`, `repoNameTrgmSimilarity`, `filePathTrgmSimilarity`, `commitHashTrgmSimilarity`, `embeddingTextTrgmSimilarity`, `searchScore`
 Fields provided by the Unified Search plugin. Includes full-text search (tsvector/BM25), trigram similarity scores, and the combined searchScore. Computed fields are read-only and cannot be set in create/update operations.
 
 ## Usage
 
 ```bash
 agentic-db document list
+agentic-db document list --where.<field>.<op> <value> --orderBy <values>
+agentic-db document list --limit 10 --after <cursor>
+agentic-db document find-first --where.<field>.<op> <value>
+agentic-db document search <query>
 agentic-db document get --id <UUID>
-agentic-db document create --entityId <UUID> --title <String> [--url <String>] [--content <String>] [--sourceType <String>] [--isRead <Boolean>] [--savedAt <Datetime>] [--parentDocumentId <UUID>] [--abstract <String>] [--overview <String>] [--activeCount <Int>] [--lastAccessedAt <Datetime>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>]
-agentic-db document update --id <UUID> [--entityId <UUID>] [--title <String>] [--url <String>] [--content <String>] [--sourceType <String>] [--isRead <Boolean>] [--savedAt <Datetime>] [--parentDocumentId <UUID>] [--abstract <String>] [--overview <String>] [--activeCount <Int>] [--lastAccessedAt <Datetime>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>]
+agentic-db document create --content <String> [--title <String>] [--metadata <JSON>] [--repoName <String>] [--filePath <String>] [--commitHash <String>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>] [--embeddingStale <Boolean>]
+agentic-db document update --id <UUID> [--title <String>] [--content <String>] [--metadata <JSON>] [--repoName <String>] [--filePath <String>] [--commitHash <String>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>] [--embeddingStale <Boolean>]
 agentic-db document delete --id <UUID>
 ```
 
 ## Examples
 
-### List all document records
+### List document records
 
 ```bash
 agentic-db document list
 ```
 
+### List document records with pagination
+
+```bash
+agentic-db document list --limit 10 --offset 0
+```
+
+### List document records with cursor pagination
+
+```bash
+agentic-db document list --limit 10 --after <cursor>
+```
+
+### Find first matching document
+
+```bash
+agentic-db document find-first --where.id.equalTo <value>
+```
+
+### List document records with field selection
+
+```bash
+agentic-db document list --select id,id
+```
+
+### List document records with filtering and ordering
+
+```bash
+agentic-db document list --where.id.equalTo <value> --orderBy ID_ASC
+```
+
+### Vector similarity search via `embedding` (manual vector)
+
+```bash
+# Pass a pre-computed vector array via dot-notation
+agentic-db document list --where.embedding.vector '[0.1,0.2,0.3]' --where.embedding.distance 1.0 --select title,embeddingVectorDistance
+```
+
+### Vector semantic search via `embedding` with --auto-embed
+
+```bash
+# --auto-embed converts text to vectors using the configured embedder (e.g. Ollama nomic-embed-text)
+EMBEDDER_PROVIDER=ollama agentic-db document search "semantic query" --auto-embed --select title,embeddingVectorDistance
+EMBEDDER_PROVIDER=ollama agentic-db document list --where.embedding.vector "semantic query" --auto-embed --select title,embeddingVectorDistance
+```
+
+### Create/update with auto-embedded `embedding` via --auto-embed
+
+```bash
+# --auto-embed on create/update converts text strings in vector fields to embeddings before saving
+EMBEDDER_PROVIDER=ollama agentic-db document create --embedding "text to embed" --auto-embed
+EMBEDDER_PROVIDER=ollama agentic-db document update --embedding "new text to embed" --auto-embed
+```
+
+### BM25 keyword search via `bm25EmbeddingText`
+
+```bash
+agentic-db document list --where.bm25EmbeddingText.query "search query" --select title,embeddingTextBm25Score
+```
+
+### Fuzzy search via trigram similarity (`trgmTitle`)
+
+```bash
+agentic-db document list --where.trgmTitle.value "approximate query" --where.trgmTitle.threshold 0.3 --select title,titleTrgmSimilarity
+```
+
+### Fuzzy search via trigram similarity (`trgmContent`)
+
+```bash
+agentic-db document list --where.trgmContent.value "approximate query" --where.trgmContent.threshold 0.3 --select title,contentTrgmSimilarity
+```
+
+### Fuzzy search via trigram similarity (`trgmRepoName`)
+
+```bash
+agentic-db document list --where.trgmRepoName.value "approximate query" --where.trgmRepoName.threshold 0.3 --select title,repoNameTrgmSimilarity
+```
+
+### Fuzzy search via trigram similarity (`trgmFilePath`)
+
+```bash
+agentic-db document list --where.trgmFilePath.value "approximate query" --where.trgmFilePath.threshold 0.3 --select title,filePathTrgmSimilarity
+```
+
+### Fuzzy search via trigram similarity (`trgmCommitHash`)
+
+```bash
+agentic-db document list --where.trgmCommitHash.value "approximate query" --where.trgmCommitHash.threshold 0.3 --select title,commitHashTrgmSimilarity
+```
+
+### Fuzzy search via trigram similarity (`trgmEmbeddingText`)
+
+```bash
+agentic-db document list --where.trgmEmbeddingText.value "approximate query" --where.trgmEmbeddingText.threshold 0.3 --select title,embeddingTextTrgmSimilarity
+```
+
+### Composite search (unifiedSearch dispatches to all text adapters)
+
+```bash
+agentic-db document list --where.unifiedSearch "search query" --select title,embeddingTextBm25Score,titleTrgmSimilarity,contentTrgmSimilarity,repoNameTrgmSimilarity,filePathTrgmSimilarity,commitHashTrgmSimilarity,embeddingTextTrgmSimilarity,searchScore
+```
+
+### Search with pagination and field projection
+
+```bash
+agentic-db document list --where.unifiedSearch "query" --limit 10 --select id,title,searchScore
+agentic-db document search "query" --limit 10 --select id,title,searchScore
+```
+
 ### Create a document
 
 ```bash
-agentic-db document create --entityId <UUID> --title <String> [--url <String>] [--content <String>] [--sourceType <String>] [--isRead <Boolean>] [--savedAt <Datetime>] [--parentDocumentId <UUID>] [--abstract <String>] [--overview <String>] [--activeCount <Int>] [--lastAccessedAt <Datetime>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>]
+agentic-db document create --content <String> [--title <String>] [--metadata <JSON>] [--repoName <String>] [--filePath <String>] [--commitHash <String>] [--tags <String>] [--embeddingText <String>] [--embedding <Vector>] [--embeddingStale <Boolean>]
 ```
 
 ### Get a document by id

.agents/skills/cli-default/references/documents-chunk.md

@@ -0,0 +1,105 @@
+# documentsChunk
+
+<!-- @constructive-io/graphql-codegen - DO NOT EDIT -->
+
+CRUD operations for DocumentsChunk records via agentic-db CLI
+
+**pgvector embedding fields:** `embedding`
+High-dimensional vector columns for semantic similarity search. Query via the Unified Search API pgvector adapter using cosine, L2, or inner-product distance. Supports chunk-aware search: set `includeChunks: true` in VectorNearbyInput to transparently query across parent and chunk embeddings, returning the minimum distance.
+
+**Unified Search API fields:** `searchScore`
+Fields provided by the Unified Search plugin. Includes full-text search (tsvector/BM25), trigram similarity scores, and the combined searchScore. Computed fields are read-only and cannot be set in create/update operations.
+
+## Usage
+
+```bash
+agentic-db documents-chunk list
+agentic-db documents-chunk list --where.<field>.<op> <value> --orderBy <values>
+agentic-db documents-chunk list --limit 10 --after <cursor>
+agentic-db documents-chunk find-first --where.<field>.<op> <value>
+agentic-db documents-chunk search <query>
+agentic-db documents-chunk get --id <UUID>
+agentic-db documents-chunk create --documentsId <UUID> --content <String> [--chunkIndex <Int>] [--embedding <Vector>] [--metadata <JSON>]
+agentic-db documents-chunk update --id <UUID> [--documentsId <UUID>] [--content <String>] [--chunkIndex <Int>] [--embedding <Vector>] [--metadata <JSON>]
+agentic-db documents-chunk delete --id <UUID>
+```
+
+## Examples
+
+### List documentsChunk records
+
+```bash
+agentic-db documents-chunk list
+```
+
+### List documentsChunk records with pagination
+
+```bash
+agentic-db documents-chunk list --limit 10 --offset 0
+```
+
+### List documentsChunk records with cursor pagination
+
+```bash
+agentic-db documents-chunk list --limit 10 --after <cursor>
+```
+
+### Find first matching documentsChunk
+
+```bash
+agentic-db documents-chunk find-first --where.id.equalTo <value>
+```
+
+### List documentsChunk records with field selection
+
+```bash
+agentic-db documents-chunk list --select id,id
+```
+
+### List documentsChunk records with filtering and ordering
+
+```bash
+agentic-db documents-chunk list --where.id.equalTo <value> --orderBy ID_ASC
+```
+
+### Vector similarity search via `embedding` (manual vector)
+
+```bash
+# Pass a pre-computed vector array via dot-notation
+agentic-db documents-chunk list --where.embedding.vector '[0.1,0.2,0.3]' --where.embedding.distance 1.0 --select title,embeddingVectorDistance
+```
+
+### Vector semantic search via `embedding` with --auto-embed
+
+```bash
+# --auto-embed converts text to vectors using the configured embedder (e.g. Ollama nomic-embed-text)
+EMBEDDER_PROVIDER=ollama agentic-db documents-chunk search "semantic query" --auto-embed --select title,embeddingVectorDistance
+EMBEDDER_PROVIDER=ollama agentic-db documents-chunk list --where.embedding.vector "semantic query" --auto-embed --select title,embeddingVectorDistance
+```
+
+### Create/update with auto-embedded `embedding` via --auto-embed
+
+```bash
+# --auto-embed on create/update converts text strings in vector fields to embeddings before saving
+EMBEDDER_PROVIDER=ollama agentic-db documents-chunk create --embedding "text to embed" --auto-embed
+EMBEDDER_PROVIDER=ollama agentic-db documents-chunk update --embedding "new text to embed" --auto-embed
+```
+
+### Search with pagination and field projection
+
+```bash
+agentic-db documents-chunk list --where.unifiedSearch "query" --limit 10 --select id,title,searchScore
+agentic-db documents-chunk search "query" --limit 10 --select id,title,searchScore
+```
+
+### Create a documentsChunk
+
+```bash
+agentic-db documents-chunk create --documentsId <UUID> --content <String> [--chunkIndex <Int>] [--embedding <Vector>] [--metadata <JSON>]
+```
+
+### Get a documentsChunk by id
+
+```bash
+agentic-db documents-chunk get --id <value>
+```

.agents/skills/cli-default/references/note.md

@@ -7,7 +7,7 @@ CRUD operations for Note records via agentic-db CLI
 **pgvector embedding fields:** `embedding`
 High-dimensional vector columns for semantic similarity search. Query via the Unified Search API pgvector adapter using cosine, L2, or inner-product distance. Supports chunk-aware search: set `includeChunks: true` in VectorNearbyInput to transparently query across parent and chunk embeddings, returning the minimum distance.
 
-**Unified Search API fields:** `contentBm25Score`, `embeddingTextBm25Score`, `contentTrgmSimilarity`, `abstractTrgmSimilarity`, `overviewTrgmSimilarity`, `embeddingTextTrgmSimilarity`, `searchScore`
+**Unified Search API fields:** `embeddingTextBm25Score`, `contentTrgmSimilarity`, `abstractTrgmSimilarity`, `overviewTrgmSimilarity`, `embeddingTextTrgmSimilarity`, `searchScore`
 Fields provided by the Unified Search plugin. Includes full-text search (tsvector/BM25), trigram similarity scores, and the combined searchScore. Computed fields are read-only and cannot be set in create/update operations.
 
 ## Usage
@@ -85,12 +85,6 @@ EMBEDDER_PROVIDER=ollama agentic-db note create --embedding "text to embed" --au
 EMBEDDER_PROVIDER=ollama agentic-db note update --embedding "new text to embed" --auto-embed
 ```
 
-### BM25 keyword search via `bm25Content`
-
-```bash
-agentic-db note list --where.bm25Content.query "search query" --select title,contentBm25Score
-```
-
 ### BM25 keyword search via `bm25EmbeddingText`
 
 ```bash
@@ -124,7 +118,7 @@ agentic-db note list --where.trgmEmbeddingText.value "approximate query" --where
 ### Composite search (unifiedSearch dispatches to all text adapters)
 
 ```bash
-agentic-db note list --where.unifiedSearch "search query" --select title,contentBm25Score,embeddingTextBm25Score,contentTrgmSimilarity,abstractTrgmSimilarity,overviewTrgmSimilarity,embeddingTextTrgmSimilarity,searchScore
+agentic-db note list --where.unifiedSearch "search query" --select title,embeddingTextBm25Score,contentTrgmSimilarity,abstractTrgmSimilarity,overviewTrgmSimilarity,embeddingTextTrgmSimilarity,searchScore
 ```
 
 ### Search with pagination and field projection