performance-review

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides instructional content and code examples for optimizing algorithms, memory, and database queries. No malicious behaviors or unauthorized system operations were detected.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to analyze project source code.
  • Ingestion points: Project directories and source files (e.g., src/).
  • Boundary markers: None explicitly defined in the instructions.
  • Capability inventory: Code analysis and performance report generation. No high-privilege tool execution is requested.
  • Sanitization: None specified for the input data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 11:48 AM
Security Audit — agent-trust-hub — performance-review