pre-pr-audit
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell tools to execute
git diff --name-only HEADfor mapping changes to tests and runsnode scripts/review-state.jsto verify the project's precommit state. These operations are within the expected scope of a PR auditing tool. - [DYNAMIC_EXECUTION]: The skill triggers the execution of a local Node.js script located at
scripts/review-state.jsor.claude/scripts/review-state.jsto obtain structured JSON data regarding the project's verification state. - [INDIRECT_PROMPT_INJECTION]: The tool ingests content from the local file system, including git diff outputs and Acceptance Criteria sections in documentation. This creates an attack surface where adversarial instructions could theoretically influence the audit's confidence score. Ingestion points: Git diff output, test file contents, and project documentation files. Boundary markers: None identified in the skill logic. Capability inventory: Broad file system access via
Read/Grep/Glob, shell execution permissions, and the ability to orchestrate sub-agents. Sanitization: No specific content sanitization or instruction-filtering logic is documented for the ingested file content.
Audit Metadata