fable-planning

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a 'Hard Rules' section that restricts the agent to read-only codebase access and explicitly forbids code edits, builds, installations, tests, or any state-changing commands.
  • [SAFE]: Evidence-based reasoning is mandatory; the agent must provide file paths and line numbers (path:line) for every claim, ensuring all analysis is grounded in the local environment.
  • [SAFE]: The framework includes a comprehensive edge-case validation checklist that specifically directs the agent to evaluate security concerns such as untrusted input boundaries and authorization checks in the design phase.
  • [SAFE]: No signs of obfuscation, remote code execution, or unauthorized network activity were found. All referenced assets and templates are local to the skill directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 05:26 PM
Security Audit — agent-trust-hub — fable-planning