memory-metadata-search

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists entirely of markdown documentation describing the search_notes tool. No executable scripts, subprocess calls, or network operations are present.
  • [INDIRECT_PROMPT_INJECTION]: The skill outlines a system that processes metadata fields from notes, which is an ingestion surface for untrusted data. 1. Ingestion points: Note frontmatter fields are ingested via the search_notes tool (as described in SKILL.md). 2. Boundary markers: No specific boundary markers or instructions to ignore instructions within metadata are provided. 3. Capability inventory: The skill only references search_notes, which is a retrieval-based tool. 4. Sanitization: The documentation specifies validation logic, including a regular expression [A-Za-z0-9_-]+ for keys and strict operator requirements, which mitigates basic injection attempts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 11:54 AM