adversarial-review

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands listed under - RUN: in the frozen-checks file. This constitutes dynamic command execution triggered by the agent's own processed findings.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from the specifications it reviews.
  • Ingestion points: The skill ingests data from the to-spec file and through repository search results.
  • Boundary markers: None; the agent lacks instructions to disregard instructions embedded in the input data during the review process.
  • Capability inventory: Includes shell command execution, repository-wide grep, and git operations.
  • Sanitization: None; input data is integrated into the validation logic and revision steps without filtering or validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 10:07 AM