The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process untrusted external data, specifically Git commit messages and Pull Request descriptions, to generate structured changelogs. This creates a surface for indirect prompt injection where malicious instructions embedded in a commit message could influence the agent's behavior during the generation process.
Ingestion points: Git commit history and Pull Request metadata (referenced in SKILL.md and references/details.md).
Boundary markers: Absent. There are no specific instructions to delimit or ignore instructions within the commit messages.
Capability inventory: The skill involves file-writing (generating CHANGELOG.md) and potential network operations (GitHub Actions, semantic-release).
Sanitization: Absent. The templates directly interpolate commit subjects and bodies into the output.

changelog-automation