cached-system-instructions
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Provides architectural guidance for hierarchical system caching to improve LLM performance.\n- [SAFE]: Includes legitimate Python code snippets utilizing the official Anthropic SDK for prompt cache control.\n- [SAFE]: The operational directives are focused on structural optimization for token efficiency rather than behavioral manipulation or safety bypasses.
Audit Metadata