blotcat-illustrations

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill is designed to ingest and process untrusted external data—such as article text, links, Notion pages, and Markdown files—which it then uses to generate visual metaphors and descriptive prompts for image generation.
  • Ingestion points: The workflow described in SKILL.md (Step 1) involves reading and extracting core viewpoints and cognitive anchors from external content provided by the user.
  • Boundary markers: The skill lacks explicit boundary markers or instructions to the model to ignore potentially malicious directives embedded within the processed article text.
  • Capability inventory: The skill possesses the capability to invoke image generation tools (imagegen) and perform filesystem write operations to the assets/ directory based on derived article metadata (e.g., <article-slug>).
  • Sanitization: There is no evidence of input validation or sanitization of the external content before it is interpolated into the {Illustration Theme} and {Core idea} variables within the prompt templates defined in references/prompt-template.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 08:13 AM
Security Audit — agent-trust-hub — blotcat-illustrations