mobx-formkit-multi-step

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill guides the creation of a data ingestion surface that processes untrusted user input through various form fields across multiple steps.
  • Ingestion points: User input is captured via form fields defined in SKILL.md (e.g., step1.firstName, step1.email, step2.street, step3.username).
  • Boundary markers: The implementation utilizes standard validation rules (e.g., required|email, min:6) to constrain input, but does not explicitly document boundary markers or instructions to ignore embedded prompts within the data fields.
  • Capability inventory: The skill includes a final submission capability via form.submit() in SKILL.md, which performs a network operation to transmit the collected data.
  • Sanitization: Validation is performed using the rules object, which provides basic structural and type checking, but specific sanitization for indirect prompt injection or cross-site scripting is not detailed in the provided snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 03:12 PM
Security Audit — agent-trust-hub — mobx-formkit-multi-step