The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use shell commands for version control and repository management tasks. Evidence: Usage of git log to gather change history and gh release commands (create, edit, view) for managing release notes on GitHub.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests untrusted data from external sources that could contain malicious instructions. 1. Ingestion points: Reads git log output, pull request descriptions, and issue comments as specified in the 'Gather' section of SKILL.md. 2. Boundary markers: Absent; there are no instructions for the agent to treat the input data as potentially adversarial or to ignore instructions embedded within the commits. 3. Capability inventory: The skill provides commands for creating and editing releases via the gh CLI, which allows the agent to publish content back to the repository. 4. Sanitization: Absent; the summarized notes are passed directly to publishing tools without validation or escaping of the ingested content.

release-notes