memory-quality-auditor

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill functions as a diagnostic utility for monitoring the health and accuracy of the agent's memory retrieval system. It analyzes internal metrics related to citation groundedness and retrieval drift.
  • [SAFE]: File operations are restricted to reading internal context and evaluation artifacts located within the project's standard memory directory (e.g., .claude/context/memory/). No sensitive user files, credentials, or system configurations are accessed.
  • [SAFE]: The implementation relies entirely on Node.js built-in modules (node:fs, node:path). There are no external package dependencies or remote code execution vectors identified.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 03:05 AM