Execution Model

Lifecycle

Every module execution follows this lifecycle:

Validate → Execute → Evidence → Result

1. Validate — Check required parameters, types, and constraints
2. Execute — Run the module logic
3. Evidence — Capture execution trace (inputs, outputs, timing, status)
4. Result — Return structured output to the caller

Execution Context

Each execution runs within a context that provides:

• Unique execution ID
• Working directory
• Environment variables
• Browser instance (if needed)
• Evidence store reference

Error Handling

Modules return structured results with a status field:

{
  "status": "success",
  "data": { ... },
  "evidence": { ... }
}

On failure:

{
  "status": "error",
  "error": "Description of what went wrong",
  "evidence": { ... }
}

Errors are always captured in evidence, making failures as auditable as successes.

Timeouts

Each module execution has a configurable timeout. If a module exceeds its timeout, execution is cancelled and an error is recorded in evidence.