The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard development tools git and gh (GitHub CLI) to retrieve project history and metadata. These operations are restricted to identifying changes since the last release and are consistent with the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill defines a process for analyzing untrusted external data such as Pull Request descriptions, commit messages, and issue contents.
Ingestion points: Pull Request and issue data retrieved via gh pr view, and commit history retrieved via git log and git show.
Boundary markers: The instructions do not define explicit delimiters for this data, but mitigate risk by mandating a "User-Centric Writing" approach where the agent must summarize and interpret changes rather than direct copying.
Capability inventory: Access to write to CHANGELOG.md and execute Git/GitHub CLI commands.
Sanitization: Risks are mitigated by the requirement for the agent to filter for "changelog-worthiness" and transform technical descriptions into user-facing impacts, which acts as an inherent sanitization step.

changelog-rules