The Agent Skills Directory

[COMMAND_EXECUTION]: The skill leverages standard developer CLI tools (git, gh, bk, vercel, flyctl) to perform repository management and CI/CD operations.
It utilizes CronCreate to establish a recurring schedule for monitoring pull request health and status.
[PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external sources, specifically pull request comments and review summaries.
Ingestion points: Fetches unresolved review threads, PR reviews, and issue-level comments through the GitHub API as defined in references/github-api.md.
Boundary markers: There are no explicit delimiters or instruction-bypass warnings defined for the ingested external data.
Capability inventory: The skill possesses the capability to modify local source code, commit changes, and execute shell commands via integrated platform CLIs as described in SKILL.md and references/ci-platforms.md.
Sanitization: The instructions do not specify any sanitization or validation of the comment content before it is used to generate fix plans.
Mitigation: The skill implements a critical security checkpoint in Phase 3 (Plan), where it requires explicit user review and approval of a generated fix plan before proceeding to the execution phase.
[EXTERNAL_DOWNLOADS]: The skill interacts with and retrieves metadata from well-known technology services and platforms, including GitHub, Buildkite, Vercel, and Fly.io, using official APIs and command-line interfaces.

babysit-pr