layered-recall

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a data ingestion architecture that represents an indirect prompt injection surface.\n
  • Ingestion points: The skill describes loading context from multiple local file paths, including ~/.claude/projects/*/memory/*.md, thoughts/CONTEXT.md, and mature-instincts.json.\n
  • Boundary markers: The architecture lacks specifications for delimiters or instructions to the agent to disregard instructions embedded within the retrieved memory content.\n
  • Capability inventory: The documentation mentions integration with functional tools such as instinct-loader, smart-memory-recall, intent-classifier, and graph-indexer.\n
  • Sanitization: There is no mention of sanitization or validation logic for the data ingested from the memory files before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 07:46 AM
Security Audit — agent-trust-hub — layered-recall