state-machines-pattern

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents patterns for ingesting external data into state machine context (e.g., FetchState in react-state-machines.md and CheckoutContext in xstate-patterns.md). This represents a standard architectural surface area where external input enters the application state.
  • Ingestion points: Context objects defined in SKILL.md, react-state-machines.md, and xstate-patterns.md designed to hold API responses and user input.
  • Boundary markers: Not explicitly defined in the provided code templates.
  • Capability inventory: The skill is purely instructional and does not provide executable scripts or tools with network/filesystem capabilities.
  • Sanitization: The examples focus on state transitions rather than data validation; sanitization of external data is left to the implementer.
  • [EXTERNAL_DOWNLOADS]: The documentation provides standard installation instructions for well-known and trusted libraries.
  • Evidence: References to npm install xstate @xstate/react and @xstate/test which are industry-standard tools for state management and testing.
  • [SAFE]: The skill uses deterministic transition tables and exhaustive union types to eliminate "impossible states," which is a defensive programming practice that enhances application security and reliability.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 11:58 PM
Security Audit — agent-trust-hub — state-machines-pattern