audit-security

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): No patterns of instruction override, safety filter bypass, or system prompt extraction were found.
  • [Data Exposure & Exfiltration] (SAFE): The skill does not access sensitive system paths (like .ssh or .aws) and performs no network operations.
  • [Remote Code Execution] (SAFE): No remote scripts are downloaded or executed; no package managers (npm, pip) are invoked.
  • [Indirect Prompt Injection] (LOW): This skill processes external code which could contain malicious instructions. However, its capabilities are restricted to searching and reporting. 1. Ingestion points: Project source code (Blade, Controllers) via grep_search. 2. Boundary markers: Absent. 3. Capability inventory: grep_search, local file reading, generating reports. 4. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:35 PM