The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes instructions to the AI to suppress its identity or tool attribution, specifically forbidding the inclusion of footers like "Generated with Claude Code" in PR descriptions.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from external sources.
Ingestion points: The skill reads data from git log, git diff, gh issue list, and gh issue view to generate PR content.
Boundary markers: There are no explicit instructions or delimiters provided to help the agent distinguish between data and potential instructions embedded in commit messages or issue descriptions.
Capability inventory: The skill can execute subprocesses via git and gh, including creating new Pull Requests with the processed content.
Sanitization: No sanitization or validation logic is present to filter malicious instructions from the ingested git or GitHub data before interpolation into the PR body.
[COMMAND_EXECUTION]: The skill utilizes shell command substitution ($(...)) and heredocs to dynamically build the gh pr create command using output from other local git and GitHub CLI operations.

create-pr