What problem does it solve? Repositories often lack clear onboarding documentation for AI agents, or existing AGENTS.md and CONTEXT.md files drift out of sync with the codebase. This Skill produces and refreshes both documents by combining automated repo exploration with an interactive question round, so agents start every session with accurate setup commands, conventions, and domain vocabulary. ## Core Features & Use Cases - Repo fact gathering: Dispatches a sub-agent to inspect manifests, lockfiles, CI config, scripts, and linter configs so users are never asked questions the repo already answers. - Interactive interview rounds: Uses numbered question batches with recommended answers to capture undocumented conventions, PR expectations, gotchas, and domain terminology. - Dual document output: Drafts AGENTS.md per a structured format (setup, build, test, conventions, gotchas) and CONTEXT.md as a domain glossary, with cross-referencing instead of duplicated content. - Living document maintenance: Reads existing files first, surfaces contradictions, and prunes stale sections on every update. - Use Case: Point it at a new monorepo to generate a root AGENTS.md covering workspace tooling plus a CONTEXT.md defining the team's domain terms before other agents start working in the codebase. ## Quick Start Ask the agent to onboard this repository by creating or updating its AGENTS.md and CONTEXT.md files.
