research-icp

Pass

Audited by Gen Agent Trust Hub on Jul 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses WebSearch and WebFetch to retrieve public research data from established platforms such as Reddit, LinkedIn, and G2. This functionality is essential to its purpose of building ideal customer profiles and buyer personas.
  • [SAFE]: Internal infrastructure is managed by a collection of Bun and Bash utility scripts (e.g., manifest-sync.ts, scaffold-eval-loop.ts, search-platforms.sh). These scripts perform standard filesystem, JSON, and query generation tasks without executing arbitrary remote code or unsanitized user input.
  • [SAFE]: A structured multi-agent workflow is implemented where a critic-agent validates all synthesized research against ten specific quality gates. These gates include verifying the integrity and attribution of quotes and ensuring evidence-based conclusions.
  • [SAFE]: No evidence of data exfiltration, credential exposure, or persistence mechanisms was detected. The skill follows established platform conventions for context persistence using a local experience substrate.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 4, 2026, 10:16 PM
Security Audit — agent-trust-hub — research-icp