cognition-strategy

Maps the cognition, strategy, and goal-hierarchy boundaries across the RPG engine's AI subsystems.

Updated Feb 8, 2026
One-click install
npx skills add https://github.com/ttnhan18062000/rpg-based-simulation --skill cognition-strategy-ttnhan18062000
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cognition-strategy
Source: https://github.com/ttnhan18062000/rpg-based-simulation/tree/main/.agents/skills/cognition-strategy
Command: npx skills add https://github.com/ttnhan18062000/rpg-based-simulation --skill cognition-strategy-ttnhan18062000

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Developers working on this deterministic RPG engine frequently misplace changes between src/cognition/ (entity self-model), src/strategy/ (bounded appraisal), and src/ai/goals/ (personality-driven scoring), because the boundary between internal self-modeling and strategic decision-making is subtle and repo-specific. ## Core Features & Use Cases - Boundary Clarification: Explains that src/cognition/ only produces the entity self-model (health, weaknesses, confidence, needs), while src/strategy/ and src/systems/strategic_systems/intelligence.py consume it for decisions. - Pipeline Documentation: Details the 4-step SelfModelUpdatePhase, the 7-stage BoundedStrategicAppraisalService with real scoring formulas, and the 4-tier goal hierarchy with interruption resistance. - Use Case: When asked to "improve entity decision-making", use this Skill to determine whether the change belongs in cognition, strategy, goals, or a domain consumer before editing code. ## Quick Start Ask where a change to entity decision-making belongs in this repo and which pipeline phase and scoring formula governs it.

Frequently Asked Questions about cognition-strategy

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
Where should entity decision-making changes go in this RPG engine?

Decision-making changes rarely belong in src/cognition/, which only builds the entity's self-model. Strategic decisions live in src/strategy/ and src/systems/strategic_systems/intelligence.py, while personality-driven goal scoring lives in src/ai/goals/.

How does the goal hierarchy work in the strategy layer?

Entities select the highest-scoring Project across four priority tiers: Survival, Biological, Social, and Economic. Switching requires the new goal score to exceed the current score plus an interruption margin defined by the entity's profile resistance.

What is the perception radius for entities in this simulation?

The perception radius is 10.0 units, consistent across domain_logic.py, view.py, and intelligence.py. A separate 15.0-unit radius exists only for cooperation candidate search and must not be conflated with perception.

Why do strategic leads lose certainty over time?

Strategic leads decay after 50 ticks without refresh via BeliefCycleSystem.decay_stale_beliefs, demoting certainty from APPROXIMATE to VAGUE to EXHAUSTED. PRECISE leads from direct observation never decay.

Which pipeline phases run cognition and strategic intelligence?

Phases 3-5 (self_model, information_belief, information_intent_execution) run early in the 39-phase AuthoritativeApplyPipeline, while phase 30 (strategic_intelligence) runs later, so strategic decisions see the tick's already-updated self-model and belief state.