clean-code-reviewer
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard
gitCLI commands (git status,git diff,git symbolic-ref,git merge-base,git diff --name-only,git diff --diff-filter=AR) to identify the review target, determine the base branch, and extract the list of changed files. This is legitimate behavior for a tool designed to review code changes. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (git diffs and codebase content), which serves as an ingestion point for potential indirect prompt injection.
- Ingestion points: Git diff output and source code files classified during the inspection phase (e.g.,
.ts,.py,.tsx,.css). - Boundary markers: The instructions provide a highly structured output format (Style, Correctness, Verdict sections) and specific finding templates, which help constrain the agent's output behavior.
- Capability inventory: Reading sibling skill files (e.g.,
../clean-typescript/SKILL.md), reading repository files, and executing git commands. - Sanitization: The instructions do not define specific sanitization or escaping mechanisms for the untrusted content being reviewed.
Audit Metadata