agent-pr-creator
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to execute git commands (git diff,git log,git status) and GitHub CLI commands (gh pr list,gh pr create). It dynamically assembles thegh pr createcommand using multi-line strings derived from commit history, which requires careful agent handling to prevent shell injection or malformed command execution. - [PROMPT_INJECTION]: The skill is subject to an indirect prompt injection surface because it reads and summarizes untrusted data from the repository's git history.
- Ingestion points: Analyzes outputs from
git log --onelineandgit diffin Phase 2 and Phase 4 of the SKILL.md workflow. - Boundary markers: Absent; there are no specific delimiters or guardrail instructions to ensure the agent ignores potentially malicious commands hidden within commit messages or code comments.
- Capability inventory: The skill has access to the
Bashtool and performs authenticated network operations via theghCLI. - Sanitization: No explicit sanitization or filtering of commit messages is mentioned before they are interpolated into the pull request title and body.
Audit Metadata