makepad-splash
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill describes the use of
cx.eval()andcx.eval_with_context()to evaluate strings as Splash code at runtime, which is a high-risk capability if the input string is not strictly controlled. - [INDIRECT_PROMPT_INJECTION]: The skill is explicitly designed for AI-assisted workflows where the agent generates Splash code. This creates a risk where untrusted data processed by the agent could influence the logic of the generated script. Ingestion points: AI-generated UI scripts and workflow automation; Boundary markers: Documentation mentions a sandboxed environment but lacks specific prompt delimitation for data processing; Capability inventory: Network access via
httpobject and script execution viaeval; Sanitization: No specific sanitization or validation methods are documented for generated code strings. - [COMMAND_EXECUTION]: The skill instructs the agent to prompt the user to execute the
/sync-crate-skillscommand if local reference documentation is missing or incomplete.
Audit Metadata