design-qa-review-skill

Warn

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to clone and install software from an external GitHub repository (github.com/ColourCloudSky/design-qa-review-skill.git) that does not match the stated author's verified infrastructure or known vendor patterns.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface because it ingests untrusted data from external sources and possesses capabilities that could be abused if the agent follows embedded instructions.\n
  • Ingestion points: Processes data from external web pages via pageUrl, mockup files via mockupUrl, and local implementation screenshots.\n
  • Boundary markers: The documentation lacks explicit safety delimiters or instructions to ignore potential commands embedded within design files or web content.\n
  • Capability inventory: Executes CLI commands (design-qa-review), makes network requests to screenshot services, and integrates with Slack for notifications.\n
  • Sanitization: No mechanisms for sanitizing or validating retrieved external content are mentioned in the skill instructions.\n- [COMMAND_EXECUTION]: The skill requires the execution of shell commands and the management of environment variables to perform its core functions, including interacting with external screenshot and messaging services.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 27, 2026, 12:35 AM
Security Audit — agent-trust-hub — design-qa-review-skill