baoyu-infographic

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided text, files, or URLs to structure infographic content, creating a potential attack surface.
  • Ingestion points: SKILL.md (Step 1: Save source content).
  • Boundary markers: Lacks formal delimiters in its prompt templates (e.g., references/base-prompt.md) to distinguish between instructions and external data.
  • Capability inventory: The skill utilizes write_file and image_generate tools.
  • Sanitization: Explicitly instructs the agent to strip credentials and secrets from outputs.
  • [COMMAND_EXECUTION]: The skill performs automated file operations and image generation tasks based on user input.
  • Tool usage: Employs write_file for analysis, content, and prompt storage.
  • Asset generation: Uses the image_generate tool to produce the final infographic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 07:28 PM
Security Audit — agent-trust-hub — baoyu-infographic