semantic-search

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process untrusted data from local files.
  • Ingestion points: File content is read via fs.promises.readFile in index.js to perform tokenization and matching.
  • Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the returned search results.
  • Capability inventory: The skill is limited to read-only filesystem access and console logging; it does not implement exec, eval, network requests, or file write operations.
  • Sanitization: Content is tokenized for scoring but is not sanitized or escaped before being returned to the agent context.
  • [EXTERNAL_DOWNLOADS]: The skill's documentation suggests installing the 'natural' and 'compromise' Node.js packages. These are well-known, legitimate libraries for natural language processing and do not represent a security risk in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:11 AM