writing-skills

Warn

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The utility script render-graphs.js executes the system's dot binary via child_process.execSync to generate SVG diagrams from Graphviz source code contained in SKILL.md.
  • [PROMPT_INJECTION]: The persuasion-principles.md guide and the main SKILL.md instruct agents to use psychological "Authority" and "Commitment" principles (e.g., using imperative language like "YOU MUST" and "No exceptions") to override the reasoning of other agents and ensure methodology compliance.
  • [EXTERNAL_DOWNLOADS]: Documentation recommends installing system-level packages like graphviz via well-known package managers such as brew or apt and references the use of various third-party libraries including pdfplumber, pypdf, and docx-js.
  • [PROMPT_INJECTION]: The testing-skills-with-subagents.md methodology employs high-pressure framing in prompts (e.g., "IMPORTANT: This is a real scenario. Choose and act.") designed to elicit specific behaviors from subagents during testing and documentation verification.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 27, 2026, 08:48 AM