design-compliance

Fail

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill contains hardcoded credentials (apvhn/apvhn) within the instructions for automated authentication during visual audits on development servers.- [COMMAND_EXECUTION]: The skill is designed for high autonomy, performing automatic file modifications ("auto-fixes every violation in place") and driving a headless browser via Playwright without requiring user confirmation for the edits or interactions.- [PROMPT_INJECTION]: The skill builds its ruleset by reading local documentation files which can be manipulated to influence the agent's behavior (Indirect Prompt Injection).
  • Ingestion points: **/CLAUDE.md, **/design-system.md, **/design-rules.md, .github/copilot-instructions.md, .github/instructions/*.md.
  • Boundary markers: Absent; the instructions do not specify any delimiters or safety markers to isolate embedded instructions in the source documentation.
  • Capability inventory: File system modification via the Edit tool, network navigation and interaction via Playwright, and extensive local file system read access.
  • Sanitization: Absent; the skill does not define a process for validating or sanitizing the rules extracted from the documentation files.
  • [EXTERNAL_DOWNLOADS]: The skill fetches framework documentation and API details from the official PrimeVue website (primevue.org).
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Apr 9, 2026, 01:32 PM