scribe
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a robust architecture for generating documentation like PRDs, SRS, and HLDs. It uses structured templates (e.g.,
reference/prd-template.md,reference/srs-template.md) and follows established industry standards (ISO/IEC/IEEE 29148:2018). - [SAFE]: The format conversion functionality uses standard, well-known CLI tools such as Pandoc, LibreOffice (
soffice), and WeasyPrint to transform documents between Markdown, Word, and PDF. These operations are restricted to document processing and do not involve untrusted code execution. - [SAFE]: The skill references established technology organizations and services (Anthropic, OpenAI, Vercel, GitHub, and various Linux Foundation-related ecosystems) for documentation standards like AGENTS.md and Spec-Kit. These references are informative and do not introduce supply chain risks.
- [SAFE]: Interaction triggers and handoff formats (e.g.,
reference/unified-spec/handoff-formats.md) utilize structured YAML payloads to communicate with other specialized agents. This is a standard coordination pattern for multi-agent systems and does not involve data exfiltration or sensitive information leakage. - [SAFE]: Operational guides like ADR (Architecture Decision Records) and Runbook writing (e.g.,
reference/adr-writing.md,reference/runbook-writing.md) promote best practices for documenting decisions and recovery procedures without performing dangerous system operations.
Audit Metadata