unity-code-review
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository, including staged code changes and documentation files.
- Ingestion points: SKILL.md reads contents from
git diff,git show, and files underdocs/,specs/, or.scratch/during its execution flow. - Boundary markers: The skill defines a review boundary conceptually, but does not implement technical delimiters or instruction-ignoring markers when passing this data to the parallel sub-agents.
- Capability inventory: The skill executes
gitcommands and spawns LLM sub-agents to process the untrusted content. - Sanitization: There is no evidence of sanitization or escaping of the ingested file content before it is interpolated into the sub-agent prompts.
- [COMMAND_EXECUTION]: The skill uses local shell commands to interact with the Git version control system.
- Evidence: It executes
git status --short,git diff --cached, andgit show :<path>to establish the code review set. These are standard operations for the skill's intended purpose of performing a code review.
Audit Metadata