fable-haiku
Pass
Audited by Gen Agent Trust Hub on Jul 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a structured task decomposition framework ("fable-mode") that emphasizes sequential staging and explicit verification.
- [SAFE]: Includes defensive instructions for file modification using
sed, specifically advising the use of word-boundary anchors (\b) to prevent accidental token corruption or unintended text replacement. - [SAFE]: Promotes a 'skeptical reviewer' mindset and requires confirmed evidence (via
grep,diff, or running code) before reporting errors, which helps prevent the hallucination of issues or the propagation of incorrect state. - [SAFE]: The skill operates entirely within the agent's provided tools (e.g., the standard Agent tool for subagents, and standard shell utilities for editing) and does not involve unauthorized network access, sensitive data harvesting, or obfuscated code.
Audit Metadata