anvil-conflict-resolution
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of markdown instructions and does not include any executable scripts, binaries, or configuration files.
- [PROMPT_INJECTION] (SAFE): Analysis of the instructions found no attempts to bypass safety filters, override core directives, or extract system prompts.
- [INDIRECT_PROMPT_INJECTION] (LOW): The skill defines a process for collecting votes from external nodes, which represents a potential ingestion point for malicious content. Evidence: 1. Ingestion points: 'votes from all committee nodes' (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: None provided (no code). 4. Sanitization: Absent. The severity is low because no capabilities are defined to be exploited via this surface.
Audit Metadata