consistency-checker-across-manuscript
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides purely instructional guidance for manuscript alignment review without any executable code, binary components, or privileged tool access. All instructions are focused on academic quality control and logical verification.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted manuscript drafts. While this represents an injection surface, the risk is effectively mitigated by the skill's lack of high-privilege capabilities such as network operations, file system writes, or command execution. The 'Hard Rules' explicitly prohibit the agent from taking actions based on insufficient evidence.
- [NO_CODE]: The skill distribution consists entirely of markdown documentation and a JSON evaluation report. There are no Python or Node.js scripts, no dependency manifest files, and no remote code download patterns identified.
Audit Metadata