The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to read and follow project-specific files like AGENTS.md, release documentation, and CI configurations to determine workflow steps. This introduces a surface where instructions embedded in these files could attempt to influence agent behavior.
Ingestion points: Files including AGENTS.md, release docs, CI configuration, and project scripts are read in the "Inspect before acting" phase.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore potential instructions embedded within the discovered documentation.
Capability inventory: The skill manages operations including git commit, git push, Pull Request creation, git merge, and repository tagging.
Sanitization: There is no explicit requirement to sanitize or validate the content of the repository-level documentation before processing.
[COMMAND_EXECUTION]: The skill coordinates standard Git operations. Risks are mitigated by the explicit requirement for user authorization before the agent performs any commit, push, PR creation, merge, or tag operation, as well as instructions to verify branch protections and deployment triggers.

git-delivery