fable-skill-creator
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements high-quality standards for agent development, emphasizing behavioral depth, structured decision rules, and modular documentation. The instructions focus on improving agent accuracy and reducing operational risks through clear activation contracts and failure taxonomies.
- [SAFE]: The 'Authoring Principles' explicitly include security safeguards, such as Principle 9 ('Zero Inlined Credentials'), which forbids embedding real API keys, passwords, or private tokens within skill artifacts.
- [SAFE]: The skill utilizes a 'Progressive Disclosure' pattern that minimizes token overhead while allowing the agent to access deep domain knowledge only when necessary, which is a best practice for managing context safely and efficiently.
- [SAFE]: The evaluation framework (BinEval) requires semantic breadth and adversarial testing to ensure that created skills are robust against shortcuts and edge cases, further contributing to the overall safety of the multi-agent ecosystem.
Audit Metadata