ask-the-council
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied questions and decisions which are subsequently passed to multiple parallel subagents for evaluation. This architectural pattern represents an ingestion point for untrusted data into an agentic workflow.
- Ingestion points: User input regarding decisions or questions passed to the council via the argument hint.
- Boundary markers: None explicitly defined to separate user input from agent instructions within the subagent context.
- Capability inventory: Spawns subagents via the Agent tool; mentions hand-offs to other tools for minimal builds and architectural updates (e.g.,
SOURCE_OF_TRUTH.md). - Sanitization: No explicit sanitization or validation of the user-provided prompt before interpolation into the subagent mandates.
Audit Metadata