fable5-judgment-orchestrator
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of markdown documentation and instructional frameworks. There are no associated scripts, binaries, or configuration files that execute code on the host system.
- [SAFE]: The documentation includes links to the official platform documentation for Anthropic and OpenAI. These references are used to provide context for model capabilities and do not involve downloading or executing unverified remote content.
- [PROMPT_INJECTION]: No malicious instructional patterns were found. The skill emphasizes evidence-based reasoning, resolving conflicts through primary sources, and following standard model safety boundaries, including a specific warning against attempting to extract private reasoning traces.
- [SAFE]: The skill defines structured templates for data ingestion and task delegation. It uses 'Evidence Packets' and 'Long-Context Packets' to establish clear boundaries for information processing, which helps mitigate risks associated with processing external data.
Audit Metadata