writing-skills

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a utility script, render-graphs.js, which uses the Node.js child_process.execSync module to call the system dot command (Graphviz) for rendering SVG diagrams from Markdown blocks. This behavior is restricted to local processing and is a standard part of the skill's visualization workflow.
  • [PROMPT_INJECTION]: The persuasion-principles.md and testing-skills-with-subagents.md files provide advanced instructions on using psychological principles such as 'Authority' and 'Commitment' to influence agent behavior. While these techniques resemble prompt injection patterns used to override safety filters, they are presented here in a meta-instructional context to improve the reliability and compliance of agent skills with procedural documentation.
  • [EXTERNAL_DOWNLOADS]: The documentation within anthropic-best-practices.md and render-graphs.js instructs users on installing well-known development tools and libraries, including system packages like graphviz and Python libraries like pdfplumber and pypdf, to support various automation tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 04:02 PM