writing-skills

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The provided utility script render-graphs.js uses execSync to run the system's dot command (Graphviz). This tool is used locally by the user or agent to render process diagrams extracted from SKILL.md into SVG format for visualization.
  • [PROMPT_INJECTION]: The skill's methodology, particularly in persuasion-principles.md and testing-skills-with-subagents.md, teaches the use of 'Instructional Pressure' and 'Authority' patterns (e.g., 'YOU MUST', 'IMPORTANT: This is a real scenario. Choose and act.'). While these resemble prompt injection techniques, they are explicitly employed as a framework for 'Discipline-Enforcing Skills' to ensure AI agents adhere to critical safety and quality protocols.
  • [EXTERNAL_DOWNLOADS]: The documentation (anthropic-best-practices.md) references several well-known and trusted libraries such as pdfplumber, pypdf, and graphviz. It provides standard installation commands using official package managers like pip, npm, brew, and apt for handling specific tasks like PDF processing and diagram rendering.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 04:20 PM