The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its reliance on external, potentially untrusted data from git commit logs.
Ingestion points: The skill reads commit messages and PR titles via git log commands in SKILL.md and references/release-notes.md to calculate version bumps and generate release notes.
Boundary markers: The instructions lack explicit delimiters or warnings to the agent to disregard instructions found within the commit data.
Capability inventory: The skill has broad permissions including file system writes (CHANGELOG.md), git tag/push operations, and GitHub release creation via the gh tool.
Sanitization: Commit message content is processed directly without filtering or escaping, relying on the agent's interpretation.
[COMMAND_EXECUTION]: The skill uses git and GitHub CLI commands to perform its release functions.
Evidence: Commands such as git tag, git push, and gh release create are executed. Security is maintained through mandatory human-in-the-loop confirmation before these commands are run.

release