memory-systems

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The analyzed files consist entirely of Markdown documentation and conceptual JSON/YAML examples. There is no executable script, binary, or configuration that triggers automated actions.
  • [SAFE]: The content describes architectural patterns (Working, Episodic, Semantic, Procedural memory) and provides guidance on integrating with third-party frameworks like Mem0 or Cognee, but does not perform the integrations itself.
  • [INDIRECT_PROMPT_INJECTION]: The skill mentions 'Context Poisoning' as a risk when noisy memory is retrieved, acknowledging the surface for indirect injection, but it does not implement any vulnerable data ingestion points itself.
  • Ingestion points: None (conceptual guide only)
  • Boundary markers: N/A
  • Capability inventory: None
  • Sanitization: N/A
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 07:06 AM