Configuration Reference

Everything in Pathfinder is driven by a single pathfinder.yaml file. This is the full annotated reference.

server

Top-level server identity and session management.

          server: name:
          my-project-docs
          # Server name, exposed in MCP metadata
          version: 1.0.0
          # Server version string
          max_sessions: 1000
          # Global max concurrent sessions (default: 1000)
          max_sessions_per_ip:
          20
          # Max concurrent MCP sessions per IP (default: 20)
          session_ttl_minutes:
          30
          # Idle timeout for active sessions (default: 30)
          session_unused_ttl_minutes:
          15
          # Idle timeout for unused sessions (default: 15)
          allowlist:
          # IPs / CIDRs that bypass max_sessions_per_ip (default: [])
          # NOTE: entries match the per-request client IP only. Behind a
            reverse proxy
          # (Railway, Fly, Nginx, etc.) the server only ever sees the proxy's
            TCP socket
          # peer as the client IP. That IP is not stable across deploys and
            should not be
          # allowlisted — set trust_proxy: true to trust X-Forwarded-For
            instead.
          # Directly-exposed servers use the socket peer.
          - "160.79.106.35"
          # Example: Anthropic Assistant crawler -
          "10.0.0.0/8"
          # Example: internal health-probe CIDR
          trust_proxy: false
          # Honor X-Forwarded-For (default: false — see warning below)
        

name — Identifies the server in MCP tool descriptions. Required.
version — Semantic version string. Required.
max_sessions — Global cap on total concurrent sessions across all IPs. When exceeded, new connections receive a 503 with a descriptive JSON body (error: "capacity_exceeded", totalSessions, maxSessions, retryAfterSeconds, contact) and a Retry-After header. A warning is logged when sessions exceed 80% of this cap. Optional, defaults to 1000.
max_sessions_per_ip — Per-IP rate limiting. Prevents a single IP from opening too many concurrent sessions. Optional, defaults to 20.
session_ttl_minutes — Idle timeout for active sessions (sessions that have invoked at least one tool). Sessions with no activity for this many minutes are cleaned up. Optional, defaults to 30.
session_unused_ttl_minutes — Idle timeout for unused sessions (sessions that connected but never invoked a tool). Designed to shed idle MCP connections quickly while keeping active sessions alive longer. Set to 15 by default to match Railway's 15-minute SSE connection hard limit. Optional.
allowlist — Array of IPv4/IPv6 addresses or CIDR ranges that bypass max_sessions_per_ip entirely. Use for trusted crawlers (e.g. 160.79.106.35 — Anthropic Assistant) or internal health-probe ranges. Optional, defaults to empty. Entries are matched against the per-request client IP — behind a reverse proxy this requires trust_proxy: true; with the default trust_proxy: false the server only sees the proxy's TCP socket address and no upstream client IP will ever match. Directly-exposed servers use the socket peer and need no extra config. Clients that hit the limit receive a 429 with a descriptive JSON body (error, reason, limit, currentCount, retryAfterSeconds, contact) and a Retry-After header.
trust_proxy — Whether to honor the X-Forwarded-For header for client-IP resolution (used by rate limiting, allowlist checks, tracing, and analytics). When true, the server populates req.ip from the leftmost entry of X-Forwarded-For. This is REQUIRED when the server runs behind a reverse proxy (Railway, Fly, Nginx, etc.) that sets X-Forwarded-For. Optional, defaults to false. ⚠️ Security: only enable this when the proxy discards any client-supplied X-Forwarded-For AND sets its own trusted value. If the proxy passes through client-supplied X-Forwarded-For, attackers can send X-Forwarded-For: 160.79.106.35 to be seen as an allowlisted IP and bypass the rate limiter entirely. When false (default), X-Forwarded-For is ignored and the server uses the TCP socket's peer address.

sources

Define where your content lives. Each source becomes a virtual filesystem that agents can explore.

          sources: - name:
          docs
          # Unique name, referenced by tools
          type: markdown
          # markdown | code | raw-text | html | document | slack | discord |
            notion
          repo:
          https://github.com/org/repo.git
          # Git repo URL (optional for local paths)
          branch: main
          # Git branch to track (default: default branch)
          path: docs/
          # Directory within the repo to index
          version: v2
          # Version tag for multi-version docs (optional)
          file_patterns:
          # Glob patterns for files to include -
          "**/*.mdx" -
          "**/*.md"
          exclude_patterns:
          # Glob patterns to exclude (optional) -
          "**/node_modules/**" -
          "**/_internal/**"
          skip_dirs:
          # Directory names to skip entirely (optional)
          - .git -
          node_modules
          max_file_size:
          100000
          # Max file size in bytes to index (optional)
          category: faq
          # Optional. Marks content as FAQ for /faq.txt and knowledge
            tools
          base_url:
          https://docs.example.com
          # Base URL for generating doc links (optional)
          url_derivation:
          # How to derive URLs from file paths (optional)
          strip_prefix:
          docs/
          strip_suffix: .mdx
          strip_route_groups:
          true strip_index:
          true chunk:
          # How to split content for embeddings
          target_tokens: 600
          # For markdown/raw-text sources
          overlap_tokens: 50
          # Token overlap between chunks
          target_lines: 100
          # For code sources
          overlap_lines: 10
          # Line overlap between chunks
        

type — Determines chunking strategy. markdown splits on headings and uses token-based chunks. code splits on function boundaries and uses line-based chunks. raw-text treats content as plain text with token-based chunks. html parses HTML structure and extracts text content with token-based chunks. document extracts text from PDF and DOCX files with token-based chunks.
category — Optional tag applied to all chunks from this source. Sources with category: faq contribute to the /faq.txt endpoint and can be queried via knowledge tools.
repo — If provided, Pathfinder clones and updates from this repo. Omit for local directories.
version — Tags all indexed chunks with this version string. Used for version-filtered search queries.
base_url — Base URL prepended to the derived slug when generating document links. Works together with url_derivation.
url_derivation — Controls how file paths are transformed into URL slugs that get appended to base_url. Each step is applied in order:
- strip_prefix — Remove this prefix from the file path before generating the URL. For example, "docs/" turns docs/guide/auth.mdx into guide/auth.mdx.
- strip_suffix — Remove the file extension. For example, ".mdx" turns guide/auth.mdx into guide/auth.
- strip_route_groups — When true, removes Next.js-style route group segments like (marketing) or (guides) from the path.
- strip_index — When true, removes a trailing /index from the path (e.g. guide/auth/index becomes guide/auth).
chunk — Use target_tokens/overlap_tokens for markdown and raw-text sources. Use target_lines/overlap_lines for code sources.

Note: The chunk config is available on all source types, but the effect varies. Markdown, raw-text, HTML, and document sources use token-based chunks (target_tokens/overlap_tokens). Code sources use line-based chunks (target_lines/overlap_lines). For Slack and Discord, the Q&A chunker produces one chunk per Q&A pair and chunk settings have no effect. For Notion, the markdown chunker respects target_tokens and overlap_tokens.

Source type: markdown

Splits content on Markdown headings and uses token-based chunks. Best for documentation written in .md or .mdx files. Configure with target_tokens and overlap_tokens.

Source type: code

Splits on function/class boundaries and uses line-based chunks. Best for source code. Configure with target_lines and overlap_lines.

Source type: raw-text

Treats content as unstructured plain text with token-based chunks. Use when content has no Markdown or code structure.

Source type: html

Parses HTML structure, extracts text content, and uses token-based chunks. Best for static sites, rendered documentation, or any HTML content.

Source type: slack

Indexes Slack threads as Q&A pairs. An LLM distills each qualifying thread into a question-answer pair with a confidence score. All pairs are stored; confidence filtering happens at query time.

          sources: - name:
          community-support
          type: slack
          category: faq
          channels:
          - C06ABC123
          # Channel IDs to index -
          C06DEF456
          confidence_threshold:
          0.7
          # Min confidence for query results (0-1)
          trigger_emoji:
          white_check_mark
          # Emoji reaction triggers immediate reindex of thread
          min_thread_replies:
          2
          # Minimum replies for a thread to be indexed
          distiller_model:
          gpt-4o-mini
          # LLM model for distillation (default: gpt-4o-mini)
        

channels — Array of Slack channel IDs to monitor. The bot must be a member of each channel.
confidence_threshold — Minimum confidence score (0-1) for Q&A pairs to appear in query results. Pairs below this threshold are still stored but filtered out at query time.
trigger_emoji — When a user reacts with this emoji on a thread, it triggers immediate reindexing of that thread. Useful for curating high-quality answers.
min_thread_replies — Threads with fewer replies are skipped during indexing.
distiller_model — The OpenAI model used to distill threads into Q&A pairs. Defaults to gpt-4o-mini.

Required environment variables: SLACK_BOT_TOKEN, SLACK_SIGNING_SECRET (for emoji-triggered webhook verification).

Source type: discord

Indexes Discord channels as Q&A pairs. Supports two channel types with different extraction strategies.

          sources: - name:
          discord-support
          type: discord
          category: faq
          guild_id:
          "1234567890"
          # Discord server (guild) ID
          channels:
          - id: "1111111111"
          # Text channel — LLM distillation
          type: text -
          id: "2222222222"
          # Forum channel — direct Q&A extraction
          type: forum
          confidence_threshold:
          0.7
          # Min confidence for query results (0-1)
          min_thread_replies:
          2
          # Minimum replies for text threads
        

guild_id — The Discord server ID.
channels — Array of channel objects, each with an id and type.
type: text — Text channels use LLM distillation (same as Slack). Threads are distilled into Q&A pairs with confidence scores.
type: forum — Forum channels extract Q&A directly from the post title (question) and replies (answer). Confidence is always 1.0 since the structure is explicit.
confidence_threshold — Minimum confidence for query results. Forum posts always pass (confidence 1.0).
min_thread_replies — Applies to text channel threads only.
distiller_model — OpenAI model for thread distillation on text channels. Optional, defaults to gpt-4o-mini.

Discord does not support emoji-triggered reindexing (requires Gateway WebSocket, which Pathfinder does not maintain).

Required environment variables: DISCORD_BOT_TOKEN, DISCORD_PUBLIC_KEY (for webhook Ed25519 verification). The bot needs the MESSAGE_CONTENT privileged intent and Read Message History + View Channels permissions.

Source type: notion

Index Notion pages and database entries as searchable markdown documents. Blocks are recursively converted to markdown, and database entry properties are serialized as YAML frontmatter.

          sources: - name:
          notion-wiki type:
          notion
          root_pages:
          - "abc123def456..."
          # Notion page IDs to index (including children)
          databases:
          - "789xyz..."
          # Database IDs — all entries indexed
          max_depth: 5
          # How deep to recurse into child pages (1-20, default: 5)
          include_properties:
          true
          # Serialize database properties as YAML frontmatter (default:
            true)
          chunk:
          target_tokens: 600
          overlap_tokens: 50
        

root_pages — Array of Notion page IDs. Each page and its children (up to max_depth) are indexed. Optional, defaults to [].
databases — Array of Notion database IDs. All entries in each database are indexed with their properties. Optional, defaults to [].
max_depth — Maximum depth for recursive child page discovery. Range 1-20, default 5.
include_properties — When true, database entry properties (Status, Priority, Tags, etc.) are prepended as YAML frontmatter to the page content. Default true.

If both root_pages and databases are empty, all pages accessible to the integration token are indexed. Requires NOTION_TOKEN environment variable.

By default, Notion sources are indexed as documents and referenced by search tools. To also make them available via knowledge tools and the /faq.txt endpoint, add category: faq to the source config — useful for Notion databases structured as Q&A or FAQ collections.

Source type: document

Index PDF and DOCX files. Requires optional peer dependencies: npm install pdf-parse for PDF support and npm install mammoth for DOCX support. Uses token-based chunks like markdown sources.

          sources: - name:
          manuals type:
          document path:
          ./documents
          file_patterns:
          - "**/*.pdf" -
          "**/*.docx"
          max_file_size:
          10485760
          # 10MB default limit
          chunk:
          target_tokens: 600
          overlap_tokens: 50
        

PDF support — Requires pdf-parse peer dependency. Text is extracted page-by-page. Scanned PDFs (image-only pages with no extractable text) are detected and skipped with a warning.
DOCX support — Requires mammoth peer dependency. Document content is converted to plain text for chunking.
max_file_size — Defaults to 10MB (10485760 bytes) for document sources. Large files are skipped with a warning.

url_derivation example

Given a file at docs/(guides)/getting-started/index.mdx with the following config:

          base_url:
          https://docs.example.com/
          url_derivation:
          strip_prefix:
          docs/
          strip_suffix: .mdx
          strip_route_groups:
          true strip_index:
          true
        

The derivation steps produce:

          # Original file path
          docs/(guides)/getting-started/index.mdx

          # After strip_prefix: "docs/"
          (guides)/getting-started/index.mdx

          # After strip_suffix: ".mdx"
          (guides)/getting-started/index

          # After strip_route_groups: true
          getting-started/index

          # After strip_index: true
          getting-started

          # Final URL = base_url + slug
          https://docs.example.com/getting-started
        

tools

Tools are what agents actually call. Four types: search, bash, collect, and knowledge.

search

Semantic search over embedded content. Requires a database and embedding config.

          tools: - name:
          search-docs
          # Tool name exposed to agents
          type: search
          description:
          Search the docs
          # Description shown to agents
          source: docs
          # Which source to search
          default_limit: 5
          # Default number of results
          max_limit: 20
          # Maximum results an agent can request
          result_format:
          docs
          # docs | code | raw
          min_score: 0.3
          # Minimum cosine similarity threshold (0-1, optional)
          search_mode:
          vector
          # vector (default) | keyword | hybrid
        

Search Modes

The search_mode field controls how search queries are executed:

vector (default) — cosine similarity on embeddings. Best for semantic/conceptual queries.
keyword — PostgreSQL full-text search (tsvector/tsquery). Best for exact terms, error codes, and technical identifiers. Does not require an embedding call at query time.
hybrid — runs both vector and keyword searches in parallel, then merges results using Reciprocal Rank Fusion (RRF, k=60). Best overall recall for mixed query types.

bash

Filesystem exploration with find, grep, cat, ls, head. Works with no database.

          tools: - name:
          explore-docs
          # Tool name exposed to agents
          type: bash
          description:
          Explore the docs
          # Description shown to agents
          sources: [docs]
          # Sources to mount (can be multiple)
          bash:
          session_state:
          true
          # Persist CWD across commands
          grep_strategy:
          hybrid
          # memory | vector | hybrid
          virtual_files:
          true
          # Expose qmd and related as virtual commands
          workspace: true
          # Enable writable /workspace directory
          max_file_size:
          100000
          # Max file size for cat output (bytes)
          cache:
          max_entries: 100
          # Max cached command results
          ttl_seconds: 300
          # Cache TTL in seconds
        

grep_strategy — memory passes grep through to bash unchanged. vector intercepts grep and runs semantic search. hybrid runs bash grep first, falls back to semantic if no results.
workspace — When true, agents can write files to /workspace/. Requires a persistent volume in production (see Deploy Guide).
session_state — When true, cd persists across tool calls within a session.

collect

Data collection tools. Agents submit structured data based on a JSON schema you define. Submitted data is stored in the local PostgreSQL database in the collected_data table, with a tool_name column and a data JSONB column containing the fields from your schema. What you do with the collected data is up to you — query it directly, build dashboards, pipe it to analytics, or export it. Pathfinder stores it; you decide how to use it.

          tools: - name:
          submit-feedback
          type: collect
          description:
          Report search quality or broken links
          response:
          Thanks for the feedback!
          schema:
          rating:
          type: enum
          description:
          How helpful was the result?
          required: true
          values: [helpful,
          somewhat,
          not_helpful]
          comment:
          type: string
          description:
          Optional details
        

knowledge

FAQ and knowledge base tool. Exposes Q&A pairs from FAQ-category sources (Slack, Discord) to agents. Supports two modes: browse (no query, returns recent pairs) and search (semantic query over pairs).

          tools: - name:
          knowledge-base
          type: knowledge
          description:
          Search community Q&A knowledge base
          sources: [community-support, discord-support]
          # Sources with category: faq
          min_confidence:
          0.7
          # Override source-level confidence threshold
          default_limit: 10
          # Default number of results
          max_limit: 50
          # Maximum results an agent can request
        

sources — Array of source names. These sources should have category: faq set.
min_confidence — Override the per-source confidence threshold for this tool. Q&A pairs below this score are filtered from results.
default_limit — Number of results returned when the agent doesn't specify a limit.
max_limit — Upper bound on results the agent can request.

When called without a query, the tool returns recent Q&A pairs (browse mode). When called with a query, it performs semantic search over the pairs.

embedding

Required when using search tools. Configures how content is embedded for vector search.

          embedding: provider:
          openai
          # openai | ollama | local
          model:
          text-embedding-3-small
          # Model name (provider-specific)
          dimensions: 1536
          # Vector dimensions (must match model)
        

Provider options

Three embedding providers are supported. OPENAI_API_KEY is only required when using the openai provider.

OpenAI (default) — uses the OpenAI embeddings API. Requires OPENAI_API_KEY.

          embedding: provider:
          openai model:
          text-embedding-3-small
          dimensions: 1536
        

Ollama — calls a local Ollama instance over HTTP. No API key needed. Requires Ollama running with the model pulled (ollama pull nomic-embed-text).

          embedding: provider:
          ollama model:
          nomic-embed-text
          dimensions: 768
          base_url:
          http://localhost:11434
          # Optional, this is the default
        

Local — runs @xenova/transformers in-process. Zero external dependencies, CPU-only. Install the optional peer dependency: npm install @xenova/transformers.

          embedding: provider:
          local model:
          Xenova/all-MiniLM-L6-v2
          dimensions: 384
        

Dimension mismatch detection

If you change embedding providers or models on an existing database, Pathfinder warns at startup when the configured dimensions don't match the existing vector index. You'll need to reindex to use the new dimensions.

indexing

Required when using search tools. Controls automatic reindexing behavior.

          indexing:
          auto_reindex: true
          # Enable daily automatic reindexing
          reindex_hour_utc:
          3
          # Hour (0-23 UTC) to run daily reindex
          stale_threshold_hours:
          24
          # Re-embed chunks older than this
        

webhook

Optional. Triggers targeted reindexing when you push to GitHub.

          webhook:
          repo_sources:
          # Map GitHub repo to source names
          "your-org/your-repo": [docs] path_triggers:
          # Only reindex when these paths change
          docs: ["docs/"]
        

Point a GitHub webhook at https://your-server/webhooks/github with the push event. Set GITHUB_WEBHOOK_SECRET to match the webhook secret.

analytics

Optional. Enables query logging and the built-in analytics dashboard at /analytics.

          analytics: enabled:
          false
          # Enable analytics (default: false)
          log_queries: true
          # Log all search queries (default: true)
          # Bearer token for /api/analytics endpoints.
          # Preferred: omit `token:` and set the ANALYTICS_TOKEN env var
            instead.
          # Inline literal also works (no ${VAR} interpolation — YAML is
            loaded as-is):
          # token: "your-secret-here"
          retention_days: 90
          # Days to retain data (default: 90)
        

enabled — When true, Pathfinder logs queries and serves the analytics dashboard at /analytics. Default false.
log_queries — When true, all search queries are logged with latency and result counts. Default true (when analytics is enabled).
token — Bearer token for authenticating requests to /api/analytics/* endpoints. Prefer the ANALYTICS_TOKEN environment variable; set analytics.token only if you need the value inline (YAML is loaded as-is — no ${VAR} interpolation). One of analytics.token or the ANALYTICS_TOKEN env var is required when analytics is enabled.
retention_days — How long to keep analytics data before automatic cleanup. Default 90 days.

The analytics dashboard provides top queries, empty result tracking, and latency metrics. API endpoints: /api/analytics/summary, /api/analytics/queries, /api/analytics/empty-queries.

Authentication

Pathfinder runs an anonymous OAuth 2.1 flow for MCP clients automatically — no config needed beyond the MCP_JWT_SECRET environment variable documented in the Deployment Guide. No user accounts, no sign-up, no dashboard — clients that perform the handshake receive a token whose subject is always anonymous. This satisfies MCP clients that require OAuth (claude.ai, newer Claude Code builds) while keeping Pathfinder a pure knowledge server.

OAuth endpoints

All standard OAuth 2.1 + dynamic client registration endpoints are served automatically:

/.well-known/oauth-protected-resource — resource metadata (RFC 9728).
/.well-known/oauth-authorization-server — authorization server metadata (RFC 8414).
/register — dynamic client registration (RFC 7591). Clients register themselves on first connect.
/authorize — authorization endpoint. Renders a one-click approve page and redirects with a code.
/token — token endpoint. Exchanges codes or refresh tokens for access tokens.
/revoke — token revocation (RFC 7009).

Bearer auth on `/mcp` and `/sse`

Both transport endpoints use opportunistic bearer authentication:

Requests without an Authorization header pass through unchanged — useful for clients that don't speak OAuth.
Requests with a valid bearer token attach the token claims to the session.
Requests with an invalid or expired bearer are rejected with 401 Unauthorized and a WWW-Authenticate challenge pointing at the resource metadata endpoint.

Clients that see the 401 challenge automatically discover the OAuth server, register, and retry with a valid token. Rotating MCP_JWT_SECRET invalidates all issued tokens at once — clients re-authenticate transparently on the next request.

Example Configs

Bash-only (no database)

The simplest setup. No API keys, no database. Agents explore docs with shell commands.

          server: name:
          my-docs version:
          1.0.0

          sources:
          - name: docs
          type: markdown
          path: ./docs
          file_patterns: ["**/*.md", "**/*.mdx"]

          tools:
          - name:
          explore-docs type:
          bash description:
          Explore project documentation
          sources: [docs]
          bash:
          session_state:
          true
        

Full stack with semantic search

Bash exploration plus RAG search. Requires Postgres and an OpenAI API key.

          server: name:
          my-docs version:
          1.0.0

          sources:
          - name: docs
          type: markdown
          repo:
          https://github.com/org/repo.git
          path: docs/
          file_patterns: ["**/*.mdx", "**/*.md"]
          chunk:
          target_tokens: 600
          overlap_tokens: 50

          tools:
          - name:
          search-docs type:
          search
          description:
          Semantic search over documentation
          source: docs
          default_limit: 5
          max_limit: 20
          result_format:
          docs

          - name:
          explore-docs type:
          bash description:
          Explore documentation with shell commands
          sources: [docs]
          bash:
          session_state:
          true
          grep_strategy:
          hybrid workspace:
          true

          embedding:
          provider: openai
          model:
          text-embedding-3-small
          dimensions: 1536

          indexing:
          auto_reindex: true
          reindex_hour_utc:
          3
          stale_threshold_hours:
          24
        

Multi-repo

Index docs from multiple repositories into a single Pathfinder instance.

          server: name:
          platform-docs
          version: 1.0.0

          sources:
          - name: api-docs
          type: markdown
          repo:
          https://github.com/org/api.git
          path: docs/
          file_patterns: ["**/*.md"] chunk: {
          target_tokens:
          600,
          overlap_tokens:
          50 } - name:
          sdk-docs type:
          markdown repo:
          https://github.com/org/sdk.git
          path: docs/
          file_patterns: ["**/*.mdx"] chunk: {
          target_tokens:
          600,
          overlap_tokens:
          50 } - name:
          sdk-source type:
          code repo:
          https://github.com/org/sdk.git
          path: src/
          file_patterns: ["**/*.ts"] chunk: {
          target_lines: 100,
          overlap_lines:
          10 }

          tools:
          - name: search-api
          type: search
          description:
          Search API documentation
          source: api-docs
          default_limit: 5
          max_limit: 20
          result_format:
          docs

          - name: search-sdk
          type: search
          description:
          Search SDK documentation and source
          source: sdk-docs
          default_limit: 5
          max_limit: 20
          result_format:
          docs

          - name:
          explore-all type:
          bash description:
          Explore all documentation and source
          sources: [api-docs, sdk-docs,
          sdk-source]
          bash:
          session_state:
          true
          grep_strategy:
          hybrid

          embedding:
          provider: openai
          model:
          text-embedding-3-small
          dimensions: 1536

          indexing:
          auto_reindex: true
          reindex_hour_utc:
          3
          stale_threshold_hours:
          24
        

Multi-version

Serve multiple versions of the same docs. Agents can filter search results by version.

          server: name:
          versioned-docs
          version: 1.0.0

          sources:
          - name: docs-v1
          type: markdown
          repo:
          https://github.com/org/repo.git
          branch: v1
          path: docs/
          version: v1
          file_patterns: ["**/*.mdx"] chunk: {
          target_tokens:
          600,
          overlap_tokens:
          50 } - name:
          docs-v2 type:
          markdown repo:
          https://github.com/org/repo.git
          branch: v2
          path: docs/
          version: v2
          file_patterns: ["**/*.mdx"] chunk: {
          target_tokens:
          600,
          overlap_tokens:
          50 }

          tools:
          - name:
          search-docs type:
          search
          description:
          Search docs (specify version parameter to filter)
          source: docs-v2
          default_limit: 5
          max_limit: 20
          result_format:
          docs

          - name:
          explore-docs type:
          bash description:
          Explore documentation filesystem
          sources: [docs-v1,
          docs-v2]
          bash:
          session_state:
          true

          embedding:
          provider: openai
          model:
          text-embedding-3-small
          dimensions: 1536

          indexing:
          auto_reindex: true
          reindex_hour_utc:
          3
          stale_threshold_hours:
          24
        