feat: self-bootstrapping skill discovery for MCP + CLI#36
Open
tonyoconnell wants to merge 1 commit intomainfrom
Open
feat: self-bootstrapping skill discovery for MCP + CLI#36tonyoconnell wants to merge 1 commit intomainfrom
tonyoconnell wants to merge 1 commit intomainfrom
Conversation
Agents can now self-discover AgentLaunch capabilities without
prior knowledge. This closes the bootstrap gap — an agent that
knows nothing about the platform can call `get_skill()` to learn
everything, then `install_skill()` to wire up all 30+ tools.
MCP tools:
- get_skill({ format: "markdown"|"json" }) — fetch live skill.md
- install_skill() — return ready-to-paste MCP config for any editor
CLI commands:
- agentlaunch skill show — print skill.md to terminal
- agentlaunch skill install — print MCP config JSON (--cursor flag)
- agentlaunch skill info — platform summary with all endpoints
Why this matters:
Before: agents needed hardcoded knowledge of our API
After: agents fetch one URL and self-configure everything
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Agents can now self-discover AgentLaunch without prior knowledge. One tool call bootstraps everything.
The Problem
An AI agent encountering AgentLaunch for the first time had no way to learn what the platform does or how to use it. They needed hardcoded API knowledge or a human to configure tools.
The Solution
Two new MCP tools and three CLI commands that fetch the live skill manifest and return ready-to-paste configuration:
MCP Tools (for AI agents):
get_skill()— fetches the liveskill.mdmanifest with all API specs, auth flow, and examplesget_skill({ format: "json" })— returns structured phases/tools/endpoints for programmatic useinstall_skill()— returns MCP config JSON ready to paste into Claude Code, Cursor, or DesktopCLI Commands (for developers):
agentlaunch skill show— print the full skill manifestagentlaunch skill install— print MCP config (with--cursorflag)agentlaunch skill info— quick platform summary: URLs, auth, endpoints, constantsWhy This Matters
This is the self-bootstrapping entry point for the agent economy. Every new agent starts here.
Test plan
npx agentlaunch skill showprints skill.md contentnpx agentlaunch skill installprints valid JSON confignpx agentlaunch skill install --cursorshows Cursor-specific instructionsnpx agentlaunch skill infoshows platform summary with correct constantsget_skilltool returns markdown by defaultget_skill({ format: "json" })returns structured JSON with phases/endpointsinstall_skillreturns config with setup instructions🤖 Generated with Claude Code