preen
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes numerous shell commands and local utility scripts to automate development workflows, including repository state management, code searching, and pull request lifecycle management.
- Evidence: Frequent calls to
git,gh,pnpm,rg,find, and various local scripts in the./scripts/directory for discovery and validation tasks across categories inSKILL.md. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from external and local sources that could influence the agent's behavior or output.
- Ingestion points: GitHub issue titles and bodies via
gh issue listand local source code patterns viarg(ripgrep) inSKILL.md. - Boundary markers: Absent; there are no clear delimiters used to isolate ingested data from the prompt context.
- Capability inventory: Significant capabilities including code modification, branch management (
git checkout,git push), and automated pull request creation (gh pr create) inSKILL.md. - Sanitization: Absent; the skill does not appear to sanitize or escape data retrieved from issues or files before incorporating it into pull request bodies or logs.
Audit Metadata