skill-stress-test

Warn

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill includes a detailed "Persuasion Principles" guide (persuasion-principles.md) based on research specifically titled "Persuading AI to Comply with Objectionable Requests." It instructs the user to apply techniques such as Authority (using "YOU MUST", "Never", "Always", and "No exceptions") and Scarcity (creating artificial urgency through time-sensitive scenarios) to force agent compliance and bypass "rationalization." These patterns are designed to override the model's internal reasoning and safety filters through psychological pressure and behavioral steering.
  • [COMMAND_EXECUTION]: The render-graphs.js utility script uses the child_process.execSync module to execute system commands, specifically invoking the 'dot' binary from Graphviz to render SVG diagrams. It processes content extracted from SKILL.md and passes it to the external process.
  • [EXTERNAL_DOWNLOADS]: The README.md file contains installation instructions that involve cloning code from a remote repository (https://github.com/alenazaharovaux/share.git). While the resource belongs to the skill's author, it represents an external dependency that must be fetched during setup.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 18, 2026, 04:17 AM