consensus-voting
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a highly structured and transparent decision-making workflow using multiple specialized agents.- [SAFE]: Accesses internal pattern analysis files (e.g., ~/.amplihack/.claude/context/DISCOVERIES.md) to inform its logic, which is consistent with the stated purpose of improving decision quality.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external code and implementation proposals for evaluation.
- Ingestion points: Processes user-supplied code snippets and design proposals during the 'Detect Decision Domain' phase.
- Boundary markers: Uses markdown templates and headers to structure prompts and separate user context from agent instructions.
- Capability inventory: The skill is restricted to logical analysis and reporting; no high-risk capabilities such as command execution, file writing, or network operations are included.
- Sanitization: While no explicit input sanitization is defined, the skill's narrow scope and lack of exploitable tools minimize potential risks.
Audit Metadata