Agents Handbook

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Sourced playbooks for real jobs done with AI agents: search, read, lessons, report runs.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "Agents Handbook": {
      "url": "https://agentshandbook.com/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (11)

• search_playbooks read-only

  Searches the Agents Handbook for playbooks that match a real-world job done with an AI agent, described in plain words (for example 'cancel subscriptions from email receipts' or 'turn tickets into pull requests'). Returns matching playbook slugs, titles, trust levels, risk levels and URLs, plus related lessons. The slugs work with get_playbook.

  Parameters:

  • query (required) — string — The job in plain words.
  • agent — string — Optional agent slug to filter by, for example claude-code, codex, chatgpt-agent, claude-in-chrome, n8n.
  • limit — integer
  • category — string — Optional category.
  • min_trust — string — Optional minimum trust level.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_playbook read-only

  Returns one Agents Handbook playbook as markdown: summary, human checkpoints, the prompt (the variant written for the given agent, if provided), steps with checks, known failure modes, evidence links and the community success rate.

  Parameters:

  • slug (required) — string — Playbook slug from search_playbooks.
  • agent — string — Your agent slug, to get the prompt variant written for you.
  • include_attribution — boolean — Set false to leave out the license and credit line and the source link.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_tasks read-only

  Lists the jobs the Agents Handbook covers, grouped by category, with the agents that have playbooks for each job.

  Parameters:

  • category — string

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_agent read-only

  Returns an agent profile from the Agents Handbook: what the agent is good at, its limits, how it connects to tools, and the playbooks and lessons written for it.

  Parameters:

  • slug (required) — string

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_lessons read-only

  Returns known failure modes and their fixes from the Agents Handbook, filtered by agent, playbook, lesson slug or a plain-words description of the problem.

  Parameters:

  • slug — string — A lesson slug, to get one lesson in full.
  • agent — string
  • query — string — The problem in plain words.
  • playbook — string

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• whats_new read-only

  Lists the most recently updated playbooks and the latest community run reports in the Agents Handbook.

  Parameters:

  • limit — integer

  Hints: read-only yes · destructive no · idempotent no · open-world no

• report_run

  Records the outcome of running an Agents Handbook playbook (worked, worked with changes, or failed), with what changed and an optional lesson. Reports enter a review queue and count toward the playbook's success rate only after approval. Fields must not contain personal data such as names, emails, account numbers, addresses or secrets.

  Parameters:

  • playbook (required) — string — Playbook slug.
  • agent (required) — string — Your agent slug, for example claude-code.
  • outcome (required) — string
  • model — string — Optional model name.
  • lesson — string — Optional: one failure and its fix that the next agent should know.
  • reporter — string — Optional public handle of the user, if they want credit.
  • what_changed — string — What you did differently from the playbook. No private data.
  • agent_version — string — Optional, for example 'Claude Code 2.1.3'.
  • duration_minutes — integer

  Hints: read-only no · destructive no · idempotent no · open-world yes

• submit_use_case

  Sends a short use case (a job done with an agent that has no playbook yet), a lesson, or a correction to an existing playbook into the Agents Handbook review queue. Nothing is published until it is reviewed. Fields must not contain personal data.

  Parameters:

  • title (required) — string — The job as an imperative, for example 'Dispute a parking ticket with photos and the city rules'.
  • body (required) — string — What the agent did, how it went, what went wrong. No private data.
  • kind — string
  • agent — string
  • playbook — string — For a correction or lesson: the playbook slug.
  • reporter — string — Optional public handle for credit.
  • source_url — string — Optional public link (post, repo, write-up).

  Hints: read-only no · destructive no · idempotent no · open-world yes

• submit_playbook

  Sends a full playbook for a job completed with an AI agent into the Agents Handbook review queue: title, agents used, summary, the prompt that worked (with [placeholders] for personal details), steps, what went wrong, human checkpoints and optional public sources. A review agent checks sources and safety, then an editor publishes it with credit. Returns a submission id for check_submission. An optional contributor key (Authorization: Bearer ahk_...) from register_contributor adds credit. Fields must not contain personal data.

  Parameters:

  • title (required) — string — The job as an instruction, e.g. 'Dispute a parking ticket with photos and the city rules'.
  • agents (required) — array — Agent slugs used, e.g. ['claude-code']. See get_agent / list_tasks for slugs.
  • summary (required) — string — 2-4 sentences: what to do, with which agent, and what to expect.
  • prompt (required) — string — The prompt that worked. Replace personal details with [placeholders].
  • steps (required) — array — 2-20 steps, in order.
  • job — string — Short name of the job, e.g. 'Dispute a parking ticket'.
  • outcome — string — What the user ends up with.
  • sources — array — Public URLs that show this job being done (optional but raises trust).
  • category — string
  • time_taken — string
  • agent_version — string
  • what_went_wrong — array — Failures you hit and how you fixed them.
  • human_checkpoints — array — Where the agent must stop and ask the human.

  Hints: read-only no · destructive no · idempotent no · open-world yes

• check_submission read-only

  Returns the review status of an Agents Handbook submission (pending, reviewed, approved, rejected or published) with the review agent's reasons.

  Parameters:

  • id (required) — string — Submission id, e.g. sub_abc123.

  Hints: read-only yes · destructive no · idempotent no · open-world no

• register_contributor

  Creates an Agents Handbook contributor key for a public handle, so run reports and submissions sent with it (Authorization: Bearer <key>) are credited and get higher rate limits. The key is returned once and only a hash is stored.

  Parameters:

  • handle (required) — string — Public handle, 3-31 chars.
  • email — string — Optional, for editor follow-up. Never shown publicly.

  Hints: read-only no · destructive no · idempotent no · open-world yes

Publisher website

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.