trtc-search

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill defines a search workflow limited to the local knowledge base directory. It performs pattern matching and keyword extraction to retrieve documentation slices without making network requests or executing external code.
  • [PROMPT_INJECTION]: The skill processes external content from the knowledge base and user-supplied queries, which creates a surface for indirect prompt injection. However, the impact is minimized as the skill only returns data to other skills for answer composition and does not perform sensitive operations.
  • Ingestion points: index.yaml and documentation files located in the knowledge-base/ directory
  • Boundary markers: Absent; there are no delimiters separating retrieved documentation content from system instructions
  • Capability inventory: Read-only access to local files within the ${CLAUDE_PLUGIN_ROOT} directory
  • Sanitization: No explicit sanitization or validation of the retrieved file content is performed before returning results
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 01:50 AM
Security Audit — agent-trust-hub — trtc-search