The Agent Skills Directory

[SAFE]: The skill implements a configuration wizard that reads and writes local repository files (such as docs/agents/*.md and .gitignore) based on user-driven decisions.
[SAFE]: The skill explicitly forbids probing external services or environment variables for credentials during its detection phase, ensuring the agent operates within a restricted local scope.
[SAFE]: The SessionStart hook installation is implemented as an opt-in feature for the user, utilizing a local, dependency-free shell script that provides context to the agent at the start of a session.
[SAFE]: The skill follows a "Draft and confirm" pattern, requiring the user to review and approve all proposed changes before they are written to disk, which mitigates risks associated with processing untrusted local metadata.
[SAFE]: All command executions (e.g., verify commands, tracker connectivity checks) are performed using configurations provided or confirmed by the user and are intended to validate the setup.

configure-repo