Interface: ToolForgeResult
Defined in: packages/agentos/src/cognition/emergent/types.ts:884
Result returned after a forge_tool invocation.
On success the new tool is registered and available immediately. On failure
the verdict field explains why the judge rejected the tool.
Properties
error?
optionalerror:string
Defined in: packages/agentos/src/cognition/emergent/types.ts:912
Human-readable error message for system-level failures (e.g., sandbox crash,
schema parse error). Distinct from judge rejection — check verdict for those.
success
success:
boolean
Defined in: packages/agentos/src/cognition/emergent/types.ts:888
true when the tool was forged, judged, and registered successfully.
tool?
optionaltool:EmergentTool
Defined in: packages/agentos/src/cognition/emergent/types.ts:899
The full emergent tool record, present only when success is true.
toolId?
optionaltoolId:string
Defined in: packages/agentos/src/cognition/emergent/types.ts:894
The assigned tool ID, present only when success is true.
Example
"emergent:a1b2c3d4-e5f6-..."
trace?
optionaltrace:ForgeTrace
Defined in: packages/agentos/src/cognition/emergent/types.ts:921
The forge's trace: what was asked, what ran, what the judge said and
where the forge ended, JSON-safe, capped and frozen. Set by
EmergentCapabilityEngine.forge on every result it returns, as the last
key. It is for the host: forge_tool takes it off the result before the
result reaches the model, so the model never reads it.
verdict?
optionalverdict:CreationVerdict
Defined in: packages/agentos/src/cognition/emergent/types.ts:906
The judge's creation verdict.
Present whether the forge succeeded or was rejected — callers can inspect
verdict.reasoning to understand why a rejection occurred.