write-manual
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation describes a workflow that involves running local Node.js and Python scripts, such as
audit_manuals.jsandcapture/run.py. These scripts are part of the project's own codebase and are used exclusively for validating documentation consistency and recording system logs in a deterministic manner. - [EXTERNAL_DOWNLOADS]: The publishing process utilizes
pagedjs-cli, a well-known open-source tool for PDF generation. This package is fetched from the official NPM registry, following standard development practices for documentation pipelines. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external technical specifications and schemas. While this represents a data ingestion surface, the framework's requirement for a restricted Markdown subset, verbatim citations verified against archived sources, and a mandatory 'TOTAL 0' audit process effectively prevents the agent from being influenced by instructions embedded within the subject material.
Audit Metadata