makepad-splash

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a standard technical documentation reference for the Makepad Splash DSL. All documented functionalities are consistent with the stated purpose of assisting with UI scripting and automation.
  • [COMMAND_EXECUTION]: The documentation describes the use of cx.eval() and cx.eval_with_context() for runtime execution of Splash scripts. These are presented as native capabilities of the Makepad framework and are common in dynamic scripting environments.
  • [EXTERNAL_DOWNLOADS]: The skill includes usage examples for a built-in http object capable of performing GET and POST requests. The examples provided use generic placeholder domains (e.g., api.example.com) for instructional purposes.
  • [PROMPT_INJECTION]: The skill contains an operational instruction for the AI to ingest content from a local reference file ./references/splash-tutorial.md. This represents a surface for indirect prompt injection.
  • Ingestion points: The agent reads from ./references/splash-tutorial.md (relative path).
  • Boundary markers: No specific delimiters or "ignore instructions" wrappers are defined for the file ingestion.
  • Capability inventory: The skill allows the agent to generate code using cx.eval (runtime execution) and http (network access) capabilities.
  • Sanitization: No sanitization of the documentation content is mentioned before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 11:33 AM