writing-skills

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a Node.js script render-graphs.js that uses child_process.execSync to execute the system dot command. This script extracts Graphviz content from the markdown documentation and spawns a subprocess to render it as an SVG file.
  • [PROMPT_INJECTION]: The skill documentation, specifically in persuasion-principles.md and CLAUDE_MD_TESTING.md, advocates for and provides examples of using 'Authority' and 'Commitment' principles to force agent behavior. This includes the use of high-pressure markers such as 'YOU MUST', 'No exceptions', and 'THIS IS EXTREMELY IMPORTANT' which are designed to override the agent's standard reasoning process in favor of strictly defined workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 04:22 PM