baoyu-cover-image

Pass

Audited by Gen Agent Trust Hub on Feb 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted text from user-provided articles to generate image prompts.
  • Ingestion points: Step 1 allows the agent to read content from arbitrary file paths or direct text input provided by the user, which is saved as source-article.md.
  • Boundary markers: Step 5 (Create Prompt File) shows that extracted information like 'Main topic' and 'Visual elements' are interpolated directly into a markdown template for the image generator without explicit delimiters or instructions to ignore embedded commands.
  • Capability inventory: The skill has the capability to write files to the local file system (cover-image/ directory) and invoke external image generation skills with the generated prompts.
  • Sanitization: There is no evidence of sanitization, filtering, or validation of the extracted content before it is used to construct the final prompt for the downstream AI tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 23, 2026, 03:10 PM