search-first

Pass

Audited by Gen Agent Trust Hub on Apr 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill's research workflow ingests data from external, untrusted sources, presenting a risk of indirect prompt injection.\n
  • Ingestion points: Data is collected from external registries (npm, PyPI, GitHub) and web searches during the 'Parallel Search' and 'Evaluate' steps.\n
  • Boundary markers: The instructions do not define delimiters or provide guidance to the agent to disregard instructions embedded in external search data.\n
  • Capability inventory: The agent has the capability to perform package installations (npm install, pip install) and file system searches (rg) based on the research results.\n
  • Sanitization: No explicit sanitization or validation of the retrieved external data is required before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 14, 2026, 12:20 AM
Security Audit — agent-trust-hub — search-first