recon

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a structured methodology for pre-build research, focusing on 'traps' (technical risks) and 'edges' (competitive advantages) using a Domain and Strategist lens.
  • [SAFE]: External communication is constrained to standard agent search tools (web search, GitHub issue exploration, and developer documentation) to gather technical insights. These operations are essential to the skill's stated purpose.
  • [SAFE]: File system interaction is limited to writing results to a localized _docs/ folder, which is a benign and expected behavior for a research-oriented skill.
  • [SAFE]: The skill emphasizes 'source quality' and requires findings to be traceable to primary sources or explicitly labeled as hypotheses, reducing the risk of hallucination or unverified claims.
  • [SAFE]: No evidence of prompt injection, obfuscation, or hardcoded credentials was found. The instructions prioritize scope calibration, ensuring the AI does not apply irrelevant or excessive constraints to small-scale projects.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 06:32 AM
Security Audit — agent-trust-hub — recon