writing-skills

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The render-graphs.js utility script executes the system's dot binary (Graphviz) to convert diagram source code into SVG files. It uses execSync to pipe content to the command's standard input, which is a common and safe implementation for local tooling.
  • [EXTERNAL_DOWNLOADS]: The anthropic-best-practices.md and SKILL.md files mention standard package installation commands for utilities like pdfplumber, pypdf, and graphviz. These instructions target official package registries and represent standard setup procedures for the tools described.
  • [PROMPT_INJECTION]: The persuasion-principles.md file outlines the use of 'Authority' and 'Commitment' language (e.g., 'YOU MUST', 'No exceptions') to influence agent behavior. While these are psychological techniques often associated with prompt engineering, they are used here to enforce strict adherence to technical documentation standards rather than to bypass safety filters or exfiltrate data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 04:39 PM
Security Audit — agent-trust-hub — writing-skills