The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes git and the gh CLI for its primary operations. It specifically instructs the agent to run network-related commands (like git push and gh pr merge) outside of the platform's sandbox to ensure connectivity to GitHub servers.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes local git status and diff data. This presents an attack surface where instructions embedded in project files could influence the agent's behavior during the grouping or summary generation steps.
Ingestion points: git status and git diff outputs processed in Step 1.
Boundary markers: No specific delimiters or safety instructions are defined to separate file content from agent instructions.
Capability inventory: git (commit, push, checkout) and gh (pull request creation, merging, branch deletion).
Sanitization: No explicit sanitization or filtering of the diff content is described.
[SAFE]: The skill implements safety rules to prevent the inclusion of sensitive files like .env or API tokens in commits and prevents irreversible actions like force-pushing.
[SAFE]: All installation and repository references point to the author's official GitHub repository (github.com/sleeplessv/relentless-data-skills).

ship