Documentation Agent

The Documentation Agent is an Agent-to-Agent (A2A) server that provides Context7-style documentation retrieval for your agents. Ask it to "look up how to set up JWT auth in Express.js" and it resolves the library name to a canonical ID, then fetches focused, version-scoped documentation - so other agents can ground their code generation in up-to-date library docs before writing against an unfamiliar API.

The agent is open-source and scaffolded with the ADL CLI. Source, releases, and the agent manifest live at github.com/inference-gateway/documentation-agent. It is published as an OCI image at ghcr.io/inference-gateway/documentation-agent.

What it does

Reach for the Documentation Agent when you want to:

• Resolve a library name to a canonical ID - turn "Next.js", "Three.js", or "Express.js" into a Context7-compatible /org/project (or /org/project/version) identifier that other tools can use.
• Fetch version-scoped documentation - retrieve up-to-date, topic-scoped docs for a specific library API, hook, configuration option, or version-specific behavior.
• Ground code generation in real docs - give other agents (like a coding agent) authoritative documentation before they write code against a third-party library, reducing hallucinated APIs.
• Skip straight to docs when you already know the ID - if you already have the canonical /org/project identifier, go directly to fetching docs without the resolution step.

It speaks the A2A protocol, so you drive it through the Inference Gateway CLI's infer agents commands, the A2A Debugger, or any A2A-compatible client.

How documentation lookup works

The agent is a Context7-compatible documentation-lookup service. A typical request is two steps: first resolve the human-readable library name into a canonical identifier, then fetch version-scoped documentation for a specific question.

Callers that already know the exact /org/project (or /org/project/version) identifier can skip straight to get_library_docs. Responses are capped at maxTokens 4096, so queries should be specific ("How to set up JWT auth in Express.js", not "auth").

Capabilities

The agent advertises the following on its A2A agent card (GET /.well-known/agent-card.json):

Capability	Value	Notes
Streaming	true	Status updates stream as the documentation is fetched.
Push notifications	false	-
State transition history	true	Full state transition history is tracked.

Skills

The agent ships one Agent Skill loaded into its system prompt, vendored as a bare scaffold and read on demand via the read tool:

Skill

What it covers

library-documentation-lookup

Fetch up-to-date documentation for a third-party library or framework before writing code against it. First resolves the library name to a Context7-compatible ID via resolve_library_id (when the caller does not already supply one), then retrieves focused, topic-scoped documentation via get_library_docs. Good for filling in unknowns about specific APIs, hooks, configuration options, or version-specific behavior.

Tools

The agent exposes two purpose-built documentation tools plus the read built-in from the ADK:

Tool	Source	Purpose	Key parameters
resolve_library_id	custom	Resolve an official library name into a ranked list of Context7-compatible library IDs.	libraryName (required), query (required)
get_library_docs	custom	Fetch up-to-date, topic-scoped documentation for a Context7-compatible library ID.	libraryId (required), query (required)
read	built-in	Read a file from disk; used to load a skill's SKILL.md body on demand.	file_path, offset, limit

The two documentation tools are implemented in Go in the agent itself; read is provided by the ADK runtime.

Services and runtime

• Server: a single Go binary (documentation-agent). documentation-agent start boots the A2A server on port 8080; --help and --version behave as expected. A multi-stage Dockerfile and the ghcr.io/inference-gateway/documentation-agent image are provided. It exposes the standard A2A endpoints: GET /.well-known/agent-card.json, GET /health, and POST /a2a.
• Documentation service: an internal service that implements the Context7-compatible documentation lookup protocol, backing both resolve_library_id and get_library_docs.
• LLM access: the agent calls an OpenAI-compatible chat-completions endpoint. Point it at the Inference Gateway (recommended) or any compatible provider via the A2A_AGENT_CLIENT_* variables.

Quick start

Register with the Inference Gateway CLI

Pull and run the image, then register it with your gateway in one step:

infer agents add documentation-agent http://localhost:8080 \
  --oci ghcr.io/inference-gateway/documentation-agent:latest \
  --run

See the A2A Integration guide for the full CLI workflow, then start chatting:

infer chat
> "Look up how to set up JWT authentication in Express.js"

Run it directly and poke it with the debugger

Run the image and point the A2A Debugger at it to exercise the protocol by hand:

# Start the documentation agent
docker run --rm -p 8080:8080 ghcr.io/inference-gateway/documentation-agent:latest

# In another shell, submit a task with the debugger
docker run --rm -it --network host \
  ghcr.io/inference-gateway/a2a-debugger:latest \
  --server-url http://localhost:8080 tasks submit "What are your skills?"

Configuration

The agent reads the standard ADK environment variables plus a small set of custom ones. The most relevant are below.

Category	Variable	Description	Default
Server	A2A_PORT	Server port	8080
Server	A2A_DEBUG	Enable debug logging	false
LLM Client	A2A_AGENT_CLIENT_PROVIDER	LLM provider (openai, anthropic, deepseek, ...)	-
LLM Client	A2A_AGENT_CLIENT_MODEL	Model to use	-
LLM Client	A2A_AGENT_CLIENT_API_KEY	API key for the LLM provider	-
LLM Client	A2A_AGENT_CLIENT_BASE_URL	OpenAI-compatible endpoint (e.g. the Inference Gateway)	-
LLM Client	A2A_AGENT_CLIENT_MAX_TOKENS	Maximum tokens for documentation responses	4096
Tools	TOOLS_READ_ENABLED	Enable the read tool (loads skill bodies on demand)	true

The defaults come from spec.config in agent.yaml; the env vars above override them at runtime. The agent's README documents the complete set of server, capability, storage, and authentication variables.

Related

• A2A Integration - protocol overview and how agents plug into the gateway
• A2A Registry - discover and publish A2A agents
• n8n Agent - a worked A2A agent with its own skill and tools
• Grafana Agent - another worked A2A agent, for Grafana dashboards and PromQL
• Browser Agent - a worked A2A agent for browser automation and web testing
• A2A Debugger - inspect and stream tasks against the agent
• Skills Catalog - how Agent Skills like library-documentation-lookup are authored and indexed
• ADL CLI - the toolchain this agent is scaffolded with
• Inference Gateway CLI - register and chat with the agent
• Repository - source, releases, and the agent manifest