agent-scoping

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs purely instructional and documentation tasks. It guides the user through a series of questions to generate an agent specification file.
  • [SAFE]: All file operations are restricted to the local filesystem, specifically reading from provided reference files and writing to docs/agent-spec.md.
  • [SAFE]: There are no network requests, external package dependencies, or dynamic code execution patterns identified.
  • [SAFE]: No obfuscation, prompt injection attempts, or privilege escalation commands were found within the instructions or templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 03:40 PM
Security Audit — agent-trust-hub — agent-scoping