fable5-model-routing

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides a framework for the agent to ingest and synthesize information from external documentation and subagent outputs, which represents an indirect prompt injection attack surface. * Ingestion points: External URLs for model documentation in references/source-notes.md and synthesized outputs from subagent lanes. * Boundary markers: The skill includes a Judgment Gate (references/judgment-gate.md) which serves as a checklist to verify evidence and identify speculation. * Capability inventory: The skill is primarily instructional and does not perform file system operations, network exfiltration, or command execution. * Sanitization: The instructions mandate that all progress and completion claims must be grounded in tool output, inspected sources, or named artifacts.
  • [EXTERNAL_DOWNLOADS]: The skill references external model documentation to inform its routing logic. * Evidence: Links to official Anthropic and OpenAI platforms are provided in references/source-notes.md to guide prompt engineering and model usage.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 05:37 AM
Security Audit — agent-trust-hub — fable5-model-routing