Infrastructure · Data Structures
SolverManifest
Typed solver identity and capability declaration
Definition
Mechanism
A typed declaration of a Solver's identity and capabilities. It contains: 1. Name/ID (unique identifier), 2. Competencies (what problem types it handles), 3. Tool Access (what actions it can perform), 4. Cost Model (expected resource usage), 5. Constraints (what it must never do). Manifests enable runtime solver selection and composition. It serves as the boot-loader for worker mode.
Invariants
- Claims, not proofs: listed capabilities are assertions the solver has not demonstrated.
- Lists the tools available to the solver, and its cost model.
- Versioned: a change of capabilities is a new version, not an edit.
- Specifies a distinct persona or system prompt.
Risks
Failure modes
Capability Inflation
Manifest claims more than solver can deliver.
Stale Manifest
Manifest doesn't reflect current solver state.
Missing Constraint
Manifest omits a critical limitation.
Relationships
Builds on
References
Non-negotiable boundary condition
Mind
Abstract five-surface contract: Manifest, Execute, Consult, Verify, Feedback
Used by
References
Mind
Partitioned feedback router for solver learning
Society
Stateful container for a running solver instance
Accepts
Society
Claim tasks and become the assigned persona
Reference
Data schema
{
"type": "object",
"description": "Typed solver identity + capabilities.",
"required": [
"solver_id",
"capabilities",
"tools",
"cost_model"
],
"properties": {
"solver_id": {
"type": "string"
},
"capabilities": {
"type": "array",
"items": {
"type": "string"
}
},
"input_types": {
"type": "array",
"items": {
"type": "string"
}
},
"output_types": {
"type": "array",
"items": {
"type": "string"
}
},
"tools": {
"type": "array",
"description": "Actions the solver can take."
},
"cost_model": {
"type": "object",
"description": "What invoking it costs."
}
}
}