skills/hubvue/skills/tech-writing/Gen Agent Trust Hub

tech-writing

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill instructions and reference materials focus on technical accuracy and editorial quality. It explicitly warns the agent to verify factual boundaries and handle sensitive data appropriately during research.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface as it is designed to ingest and process external repository content.
  • Ingestion points: Project source code, configuration files, and manuscripts (SKILL.md, references/evidence-and-artifacts.md).
  • Boundary markers: The skill relies on 'factual review gates' rather than explicit prompt delimiters for ingested content (references/review-gates.md).
  • Capability inventory: The skill utilizes file system reading for research and file writing for persisting approved manuscript sections.
  • Sanitization: Instructions in 'references/evidence-and-artifacts.md' specifically mandate checking that 'sensitive data has been handled appropriately'.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 07:13 AM
Security Audit — agent-trust-hub — tech-writing