specialist-review

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): The skill uses persona adoption for domain-specific analysis, which is its intended purpose. No instructions to bypass safety filters or ignore system prompts were found.
  • Data Exposure & Exfiltration (SAFE): The skill reads project files but does not access sensitive system paths (like ~/.ssh) or transmit data to external endpoints. Operations are confined to the local project structure.
  • Indirect Prompt Injection (SAFE): The skill ingests untrusted data from targets being reviewed.
  • Ingestion points: Step 3 uses Read, Glob, and Grep tools to examine target file content.
  • Boundary markers: Not explicitly defined in the logic to separate target content from agent instructions.
  • Capability inventory: Read, Write, Glob, and Grep tools are available to the agent.
  • Sanitization: Step 1 references input validation and sanitization for roles and targets. Since the primary purpose is code analysis, this surface is considered a functional requirement and managed by guidance.
  • Command Execution (SAFE): No unauthorized shell commands or arbitrary code execution patterns were found. The use of 'cat' in documentation is illustrative of reading an internal asset.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 03:14 PM