The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of shell commands (git, gh, bun, awk, etc.) to automate repository management, filesystem modifications, and GitHub API interactions throughout the release process.
[DYNAMIC_EXECUTION]: In Step 1.5, the skill compiles the project's source code into a temporary binary and executes it with the --help flag. This involves the execution of code generated at runtime, which poses a risk if the source code being released contains malicious module-initialization logic or side effects.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources to drive its logic.
Ingestion points: Data is read from commit messages via git log, PR summaries and titles via gh pr view, and existing CHANGELOG.md content.
Boundary markers: The skill does not employ specific delimiters or warnings to ignore instructions embedded within the ingested PR or commit data.
Capability inventory: The skill possesses significant capabilities, including the ability to write files, commit and push to remote git repositories, and create/modify GitHub releases and pull requests.
Sanitization: There is no evidence of sanitization or escaping of the ingested data before it is processed by the AI to draft changelogs and PR bodies.

release