rust-code-review
Pass
Audited by Gen Agent Trust Hub on Aug 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including git diffs, PR bodies, and issue descriptions. These are interpolated into prompts for parallel sub-agents, creating a surface where an attacker could embed malicious instructions (e.g., in a code comment or PR description) to override the reviewer's logic or conceal defects.
- Ingestion points: The sub-agent prompts in
REVIEW-PASSES.mdexplicitly ingest 'The diff' and 'The request the change answers' (issue/PR body). - Boundary markers: The prompts do not use distinct delimiters or escaping to isolate the reviewed content from the agent's instructions.
- Capability inventory: The skill executes shell commands via
cargo fmt,cargo clippy, andcargo testas part of its verification workflow. - Sanitization: No sanitization or filtering of the input data is performed before it is passed to the LLM.
Audit Metadata