performance-optimization

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWNO_CODE
Full Analysis
  • [General Security] (SAFE): No malicious code, exfiltration patterns, or obfuscation were detected. The file defines behavioral principles and expertise for an AI agent without providing any underlying scripts or tools.
  • [Indirect Prompt Injection] (LOW): The skill is intended to process user-provided performance bottlenecks. While it lacks explicit boundary markers for untrusted input, the skill itself possesses no executable capabilities that could be exploited. Evidence: 1. Ingestion points: User-provided performance descriptions and code snippets. 2. Boundary markers: Absent. 3. Capability inventory: None (informational only). 4. Sanitization: Absent.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 01:46 AM