aipixelperfect-design-engine

Warn

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The installation instructions direct the user to clone a repository from an untrusted GitHub account (abnormal-codex/Ai-Pixel-Design-Archive.git) which does not correspond to the verified vendor or author identity.
  • [REMOTE_CODE_EXECUTION]: By instructing users to run npm install and npm start on a cloned repository from an untrusted source, the skill facilitates the execution of arbitrary code through package scripts or repository source content.
  • [COMMAND_EXECUTION]: The skill documents numerous shell commands for environment setup, build processes, and CLI interactions, which can manipulate system resources and perform network operations.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface by passing user-supplied text prompts into image synthesis functions and cultural adaptation logic without proper sanitization or boundary delimiters.
  • Ingestion points: Prompts enter the system via the prompt parameter in generateDesign and advancedSynthesis functions in SKILL.md.
  • Boundary markers: The skill does not use delimiters or instructions to the model to ignore embedded malicious instructions within the prompts.
  • Capability inventory: The platform performs network requests to external APIs and uses WebSockets for real-time collaboration.
  • Sanitization: Although copyright analysis is mentioned, there is no evidence of escaping or validation for incoming prompt data to prevent behavioral overrides.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 30, 2026, 08:39 PM
Security Audit — agent-trust-hub — aipixelperfect-design-engine