Infrastructure · Data Structures
Task
The atomic unit of intent
Definition
Mechanism
A specification of intent to be performed by a System. It encapsulates what to achieve without prescribing the method (how). A Task is recursive: it can be decomposed into child Tasks. Tasks form a Hierarchy where child Constraints must be a superset of parent Constraints (Holographic Inheritance).
Invariants
- Holographic Inheritance: Constraints(Child) must include ALL Constraints(Parent).
- Intent Preservation: Decomposition must not lose the original goal.
- Linkage: Every non-root Task must reference its parent.
Risks
Failure modes
{{constraint}} Escape
Child task violates parent safety constraints.
Orphan Task
Task created without linkage to parent (except Root).
Zombie Task
Task continues after its Context is invalidated.
Ambiguous Intent
Task description admits multiple incompatible interpretations.
Relationships
Builds on
References
Non-negotiable boundary condition
Portable execution environment with inherited constraints
Vertical rank ordering with Upper/Lower relations, inheritance, and authority
Complex whole of interacting parts
Used by
References
Society
Standard patterns for multi-agent coordination
Society
Binding offer from solver to execute task
Society
Economically constrained task
Mind
Economic governor for cognitive spend
Society
Ingest only what is strictly necessary
Mind
Postponing decisions until more information available
Mind
Calculating the economic value of reducing uncertainty
Standardized learning feedback packet
Mind
Expansion of cognitive capability through recursive decomposition of concepts into contract-bounded sub-concepts
Mind
Practise on tasks just beyond what can be done alone, with support that is withdrawn as skill grows
Mind
Use prompts to generate or refine other prompts
Physics
Irrelevant information
Mind
Ordering tasks by importance and urgency
A state gap with a cost of inaction
Mind
Self-location awareness within the task graph
Ordered container with FIFO/LIFO/priority discipline and explicit dequeue semantics
Society
Continuous ownership of a system invariant
Mind
Apex triage node of a SolverTree, with Pathway Memory as its compounding learning site
Mind
Lock requirements to prevent scope creep
Society
System health acting as an inhibitory emotion
Formal state machine governing task progression
Society
Discover and invoke external tools via structured capability manifests
Execute external tool and observe result
Society
Meaning emerges from successful coordination
Society
Claim tasks and become the assigned persona
Extends
Society
Economically constrained task
Accepts
Mind
Dividing complexity into manageable units
Society
Work handoff: delegate → accept/refuse → heartbeat → result-or-failure
Society
Explicit safety reasoning prior to execution
Mind
Predictive resource costing
Mind
Scale inference compute via extended reasoning
Society
Transfer control and context between agents
Pause for human approval before critical actions
Mind
Solve subproblems from easiest to hardest sequentially
Mind
Run subtasks simultaneously and aggregate
Mind
Solver implementing the five-surface contract (polymorphic across substrates)
Society
Sequential LLM calls with validation gates
Mind
Interleaved thought-action-observation cycle
Mind
Self-critique and retry after failure
Classify input and direct to specialized handler
The fractal unit of completed work
Mind
Abstract five-surface contract: Manifest, Execute, Consult, Verify, Feedback
Society
Active hierarchy of coordinated solver instances
Rate-limiting to prevent resource exhaustion
Composes with
The application of effort to reduce entropy
Reference
Data schema
{
"type": "object",
"required": [
"operation",
"inputs",
"acceptance_criteria"
],
"properties": {
"operation": {
"type": "string",
"description": "What the Task does — the intent"
},
"inputs": {
"type": "object",
"description": "Typed inputs to the operation"
},
"acceptance_criteria": {
"type": "array",
"items": {
"type": "string"
},
"description": "Conditions that mark the Task complete; closure signal for the caller"
},
"budget": {
"type": "object",
"description": "Optional resource ceiling (compute, time, tokens)"
},
"parent_id": {
"type": "string",
"description": "Super-task reference if hierarchical"
},
"deadline": {
"type": "string",
"format": "date-time"
}
}
}