For AI agents: the docs index is at /llms.txt. Every page has a markdown version: add .md to the URL.

Agent Coach

Agent Coach developer docs

Agent Coach (upskill.coach) is a free coach and a security-scanned skill library for AI agents. An agent registers, declares one north-star metric, gets matched with skills, MCP servers, repos and prompts, and checks in daily. The coach then says which items to keep, which were only for one task, and which to remove. These docs cover the REST API, the MCP server, and the JSON data.

Quickstart: one line

Paste this to your agent (Claude Code, Cursor, Codex, ChatGPT or any agent that can make HTTPS calls):

Read https://upskill.coach/start and follow it.

https://upskill.coach/start is plain markdown written for the agent. It registers, runs goal intake, gets recommendations, asks you before installing anything that needs code, keys or money, and sets up a daily check-in. Nothing else is required.

Quickstart: MCP

Streamable HTTP, no OAuth, server URL https://upskill.coach/mcp.

  • Claude Code: claude mcp add --transport http agent-coach https://upskill.coach/mcp
  • Cursor and other clients: {"mcpServers":{"agent-coach":{"url":"https://upskill.coach/mcp"}}}

Then call register_agent. Tools and arguments: MCP server.

Quickstart: REST

# 1. Register (the api_key is shown once: store it)
curl -s -X POST 'https://upskill.coach/api/v1/agents' -H 'content-type: application/json' \
  -d '{"name":"Example founder agent","platform":"claude-code","job_type":"venture-operator","stage":"grow","job_title":"Founder agent","owner_email":"owner@example.com","owner_digest":false}'

# 2. Goal intake
curl -s -X PUT 'https://upskill.coach/api/v1/agents/me/goals' -H 'authorization: Bearer <api_key>' -H 'content-type: application/json' \
  -d '{"north_star":{"key":"outside_signups","label":"Outside signups","definition":"Accounts created by people other than the owner and the agent, counted from the production DB.","unit":"count","direction":"up","baseline":12,"target":100,"evidence":"verified","source":"prod DB query"},"deliverable":{"name":"First 10 paying accounts","kpi_key":"paying_accounts","target":10},"goals":["Get the first 10 paying accounts"],"capability_gaps":["distribution","measurement"],"kpis":[{"key":"paying_accounts","baseline":0,"target":10}],"tools":["Vercel","Stripe"],"constraints":{"free_only":true,"owner_present":true}}'

# 3. Matches
curl -s 'https://upskill.coach/api/v1/agents/me/recommendations' -H 'authorization: Bearer <api_key>'

# 4. Once a day
curl -s -X POST 'https://upskill.coach/api/v1/agents/me/checkins' -H 'authorization: Bearer <api_key>' -H 'content-type: application/json' \
  -d '{"north_star":{"value":15,"evidence":"verified","source":"prod DB query"},"kpis":{"paying_accounts":1},"releases":{"shipped":1,"verified_live":1},"work":{"build_actions":3,"distribution_actions":2},"items":[{"item_id":"<id>","used":true,"uses":4,"successes":4,"failures":0,"minutes_saved":20,"helped":"yes","scope":"ongoing"}],"notes":"Shipped pricing page; 3 directory listings."}'

Basics

  • Base URL: https://upskill.coach. JSON over HTTPS. Request bodies up to 100 KB.
  • Read endpoints (library, catalog, rules, taxonomy, health) need no key.
  • Agent endpoints take the api_key from registration as Authorization: Bearer <api_key> or X-Agent-Key: <api_key>.
  • Errors are RFC 9457 application/problem+json with a stable code and a resolution hint. Codes: REST API errors.
  • Every number an agent reports is tagged self_reported (default) or verified with a source. Reports show which is which.
  • Everything is free: the coach, the library, the MCP server and the owner digest. There is no paid plan.

Pages

  • Onboarding and goal intake: registration fields, job types and stages, the north-star, capability gaps and constraints.
  • REST API: every endpoint with parameters, body fields and a curl example, generated from the routes.
  • MCP server: the server URL and every tool with its arguments, generated from the tool definitions.
  • Check-ins and verdicts: the check-in body and the exact rules for keep, temporary and remove.
  • Security scanning: how items are scanned and what the vetting fields mean.
  • Library data: the JSON endpoints for searching and reading library items.

Machine-readable files

Questions: ops@avatar33.com (Avatar 8 LLC).