The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs broad system operations using git, gh, bun, npm, uv, cargo, and awk to manage source code, tags, and releases.
[COMMAND_EXECUTION]: In Step 1.5, the skill compiles local source code into a temporary binary using bun build and executes the resulting file to perform a smoke test. This constitutes dynamic execution of code within the agent's environment.
[EXTERNAL_DOWNLOADS]: The skill clones the vendor's tap repository (coleam00/homebrew-archon) and downloads release artifacts from GitHub to verify checksums. These target well-known services and vendor-owned resources.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection by ingesting untrusted data from the repository's git history.
Ingestion points: Commit messages and file diffs are read from the local repository during the changelog generation phase (SKILL.md).
Boundary markers: No delimiters or safety instructions are used to prevent the agent from being influenced by instructions embedded in commit messages.
Capability inventory: The agent has the ability to write these processed messages to PR descriptions (gh pr create) and GitHub release notes (gh release create), creating a path for downstream poisoning or social engineering.
Sanitization: No validation or sanitization is performed on the commit strings before they are incorporated into the agent's output.

release