cost-latency-optimizer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill handles untrusted user prompts through functions like optimize_prompt and select_model_by_complexity to reduce costs and latency. This represents an indirect injection surface. Evidence: 1. Ingestion points: optimize_prompt and select_model_by_complexity in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Local logic processing and Redis caching. 4. Sanitization: Absent.
- [SAFE] (SAFE): No malicious patterns, obfuscation, or data exfiltration attempts were detected. The skill uses standard libraries for hashing and performance monitoring.
Audit Metadata