writing-skills

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains multiple documents (e.g., testing-skills-with-subagents.md, persuasion-principles.md) that teach and utilize 'pressure scenarios' to enforce agent compliance. These scenarios employ role-play framing and override markers (e.g., "IMPORTANT: This is a real scenario. Choose and act.") which are characteristic of prompt injection techniques.
  • [COMMAND_EXECUTION]: The utility script render-graphs.js utilizes child_process.execFileSync to execute the system dot binary (Graphviz) for transforming embedded Graphviz syntax into SVG diagrams.
  • [DYNAMIC_CONTEXT_INJECTION]: The SKILL.md file provides educational guidance on the !command syntax used by some platforms to execute shell commands at skill load time and inject their output into the prompt context. While documented for legitimate development workflows, this feature represents a significant attack surface for dynamic context injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 06:18 AM
Security Audit — agent-trust-hub — writing-skills