context-optimization
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [Prompt Injection] (SAFE): No instructions found that attempt to bypass safety filters, override system constraints, or extract system prompts.
- [Data Exposure & Exfiltration] (SAFE): No sensitive file paths (~/.aws, ~/.ssh), hardcoded credentials, or unauthorized network operations detected.
- [Obfuscation] (SAFE): No hidden or encoded content (Base64, zero-width characters, homoglyphs) found.
- [Unverifiable Dependencies & RCE] (SAFE): No package manifests or commands to download and execute remote scripts were identified.
- [Indirect Prompt Injection] (INFO): The skill identifies data ingestion surfaces like tool outputs and retrieved documents, but as a documentation-only resource ('NO_CODE'), it lacks any executable capability that could be exploited.
- [Dynamic Execution] (SAFE): Code snippets are illustrative pseudocode and do not utilize unsafe functions like eval() or exec() for runtime code generation.
Audit Metadata