writing-skills

Warn

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains comprehensive documentation on how to use persuasion techniques to ensure AI compliance and bypass alignment-driven 'rationalizations.' Evidence: persuasion-principles.md details the use of 'Authority' (utilizing markers like 'YOU MUST', 'Always', and 'No exceptions') and 'Scarcity' to eliminate decision fatigue and force compliance. Evidence: testing-skills-with-subagents.md describes the use of 'pressure scenarios' involving economic, time, and authority constraints specifically designed to make agents want to violate their rules.
  • [COMMAND_EXECUTION]: The included utility render-graphs.js executes system-level commands to process data. Evidence: The script utilizes Node.js child_process.execSync to run the dot (Graphviz) command on content extracted from markdown files. Evidence: The script also executes which dot to verify the presence of the Graphviz binary on the host system.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 25, 2026, 08:24 PM