krt-interface-inquisitor

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it is designed to analyze and act upon content from external, untrusted sources.
  • Ingestion points: The skill instructions in SKILL.md direct the agent to "inspect the actual page in a browser" and process user-provided "DOM/component structure" and "CSS/layout choices."
  • Boundary markers: There are no boundary markers or instructions to treat the analyzed content as untrusted data or to ignore any embedded commands within the UI being critiqued.
  • Capability inventory: The skill generates an "Implementation brief" with concrete instructions that are consumed by other agents or implementation workflows, potentially allowing malicious instructions in a analyzed UI to influence downstream actions.
  • Sanitization: No sanitization, filtering, or validation of the retrieved browser content or DOM strings is described in the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 06:55 AM
Security Audit — agent-trust-hub — krt-interface-inquisitor