dual-round-review
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust and secure workflow for auditing code and documents. It utilizes isolated sub-agent contexts to prevent persistent instruction override and ensures the integrity of review outputs through cryptographic hashing (SHA256) and write-once protection for archived reports.
- [COMMAND_EXECUTION]: The skill executes local Git commands via Python scripts to extract diffs and history. These commands are executed as lists of arguments, preventing shell injection. Inputs such as slugs and paths are validated (e.g., kebab-case enforcement) to further mitigate misuse.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from external sources such as Git changes and requirement specifications. This attack surface is mitigated by the skill's architecture, which uses 'Context Isolation' for reviewing sub-agents and assigns them explicit adversarial 'Red-Team' personas, ensuring that embedded instructions in the ingested data do not influence the agent's behavior.
Audit Metadata