validate-tests

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection because it processes untrusted test code and feature descriptions.
  • Ingestion points: Test code and feature artifacts are ingested from user input.
  • Boundary markers: Absent. No markers are used to delimit external data.
  • Capability inventory: The skill uses vendor tools (mcp__helpmetest-*). No arbitrary command execution or file-system writes are defined in the markdown.
  • Sanitization: Absent. The skill does not sanitize or validate external input before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 04:54 AM
Security Audit — agent-trust-hub — validate-tests