agent-properties-triage

name	agent-properties-triage
description	Decision procedure for determining whether an agent property is relevant, conditionally relevant, or not relevant in the current Logistics simulation domain.

Agent Property Relevance Triage Skill

Purpose

Use this skill to quickly classify property relevance before proposing model changes or conceptual summaries.

Inputs

• User question.
• Text sections in index.html.
• Candidate board IDs involved in the request.
• Behavior tree context from scripts/Entities/Pawn.ts and scripts/Helper/Decision/Case.ts.
• Tension framework context from simulation-concepts.instructions.md (division of labour, specialization, autonomy, vision, flexibility).

Triage Procedure

1. Identify the section-level claim in index.html that the user question targets.
2. Map the claim to one or more properties from agent-properties-map.
3. Check whether requested boards are active-domain or deferred-domain.
4. Check whether the claim depends on behavior-tree policy:

• execution order,
• selector/sequence/inverter semantics,
• movement gating,
• sharing/feeding/task switching.

5. Check which organizational tension axes are implicated:

• division of labour vs autonomy,
• specialization vs flexibility,
• organizational vision vs local optimization,
• stability vs learning speed.

6. Classify each property as:
   • Relevant: directly tested or explained by active-domain boards.
   • Conditionally relevant: depends on parameter regime or scenario setup.
   • Not relevant: outside active-domain narrative or on deferred boards.
7. Return relevance with one-line rationale and board evidence.

Active vs Deferred Board Sets

• Active-domain boards:
  • trial-canvas-0
  • trial-canvas
  • trial-canvas-2
  • resource-canvas-1
  • resource-canvas-2
  • resource-canvas-3
  • knowledge-reproduction-canvas-1
  • knowledge-reproduction-canvas-2
  • knowledge-reproduction-canvas-3
  • knowledge-reproduction-canvas-4
  • knowledge-reproduction-canvas-5
  • sequence-canvas-1
  • sequence-canvas-2
  • sequence-canvas-3
  • command-canvas-1
• Deferred-domain boards:
  • hierarchy-canvas-1
  • hierarchy-canvas-2
  • hierarchy-canvas-3
  • competition-canvas-1
  • competition-canvas-2
  • competition-canvas-3
  • roles-canvas-1
  • roles-canvas-2
  • roles-canvas-3
  • democracy-canvas-1
  • democracy-canvas-2
  • democracy-canvas-3
  • reputation-canvas-1
  • reputation-canvas-2
  • reputation-canvas-3
  • market-canvas-1
  • market-canvas-2
  • market-canvas-3
• Non-domain board:
  • planets-background-canvas

Relevance Heuristics

• Relevant if the text claim and board behavior align in the same section narrative.
• Relevant for BT analysis when runtime policy nodes directly influence the claimed behavior in the active board.
• Conditionally relevant when toggles/thresholds alter behavior class:
  • giving-threshold,
  • request-threshold,
  • actors-frequency,
  • actors-pause-duration,
  • actors-pause-interval,
  • noise-scale.
• Conditionally relevant for BT when one of these changes branch outcomes or execution order effects:
  • shares,
  • consumes,
  • pauseInterval,
  • pauseDuration,
  • hungerRate,
  • movement speed,
  • knowledge freshness.
• Conditionally relevant for organizational tensions when one of these changes tradeoff balance:
  • actor count,
  • communication reliability,
  • task prerequisites,
  • handoff volume,
  • delegation strictness,
  • adaptation cadence.
• Not relevant when any of these hold:
  • board is deferred-domain,
  • question is about physics-only background canvas,
  • claim requires governance forms not currently in active scope,
  • claim concerns non-runtime prose only and does not depend on pawn decisions.

Required Response Shape

For each property, output exactly these fields:

• property
• relevance
• why
• bt_mechanism
• tension_axis
• tension_tradeoff
• boards
• not_relevant_when

Done Criteria

• Every claimed property is tied to at least one active board or explicitly marked out of scope.
• Deferred boards are not used as primary evidence.
• Non-relevance conditions are explicit and testable.
• If BT is in scope, cited mechanism includes at least one of: execution order, node semantics, gating condition, or task-switch behavior.
• If methodology tension is in scope, output includes both sides of at least one axis and identifies what is being traded off.