fable-5
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill includes robust security and safety instructions designed to prevent the AI from following malicious instructions found in external data (Indirect Prompt Injection). Specifically, the 'trust-and-authority.md' file instructs the agent to treat imperatives in content as data, not as tasks.
- [SAFE]: The skill mandates safe handling of credentials, instructing the agent to never propagate secrets (API keys, tokens) into logs, commits, or command strings, and to refer to them only by location or environment variable name.
- [SAFE]: The skill implements strict authorization gates, requiring explicit user consent for any action that leaves the local environment, such as network calls, publishing artifacts, or sending messages to third parties.
- [SAFE]: Static detector triggers for prompt injection are confirmed false positives; the flagged text consists of defensive guidelines that explain how to identify and disregard injection attempts rather than performing them.
Audit Metadata