web-research

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses WebSearch and WebFetch tools as intended for its primary purpose of online research. All network operations are directed towards fetching content for the user's requested topic.
  • [SAFE]: The skill maintains transparency by printing reports to the chat and saving them to the local ~/research-reports/ directory. This is standard behavior for a research agent.
  • [SAFE]: The practice of appending to ~/.claude/skills/web-research/LEARNINGS.md is a self-improvement mechanism (feedback loop) and does not involve sensitive data exposure or persistence-based attacks.
  • [SAFE]: The skill uses structured templates and explicit citation requirements, which mitigates risks related to hallucination or misinformation in its output.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 10:35 AM
Security Audit — agent-trust-hub — web-research