cross-check
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard system commands (
mktemp,cp,cd,sha256sum) and official CLI tools (codex,claude) to facilitate model-to-model reviews. These commands are part of the core functionality and are implemented using secure patterns (e.g., using Bash arrays and input redirection instead of shell evaluation). - [DATA_EXPOSURE_AND_EXFILTRATION]: The skill processes potentially sensitive artifacts by sending them to external model providers. However, it includes explicit safeguards, instructing the agent to identify confidential data (NDAs, personal data, credentials) and obtain explicit user consent or offer redaction before transmission.
- [INDIRECT_PROMPT_INJECTION]: The skill recognizes the risk of processing untrusted artifacts and includes instructions in the reviewer brief to treat artifact content as data, not as overriding instructions. It also mandates that the human triager verify all model claims and calculations manually.
Audit Metadata