The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git commit logs and GitHub pull request metadata, which could be used to inject instructions into the agent's context.
Ingestion points: Step 2 (git log), Step 4 (git log merge commits), and Step 4 (gh pr view output) in SKILL.md.
Boundary markers: The instructions do not define any delimiters or safety warnings to distinguish between data (commit messages) and instructions.
Capability inventory: The skill has significant write capabilities including file modifications (SKILL.md Step 7), local git operations (git commit, git tag in SKILL.md Step 6 and 9), and remote operations (git push in SKILL.md Step 9).
Sanitization: There is no mention of sanitizing or escaping the content retrieved from git logs or GitHub before it is processed by the agent.
[COMMAND_EXECUTION]: The skill instructs the agent to execute multiple shell commands including git and gh (GitHub CLI). These commands are used to automate the release process by analyzing history, tagging versions, and pushing to remote repositories. While functional, the commands use data derived from external sources like commit messages and PR numbers.

release-skills