fpf

name	fpf
description	First Principles Framework - structured reasoning skill for any task requiring auditable thinking, evidence chains, systematic problem-solving, or holonic composition. Use FPF patterns to guide reasoning on engineering, research, and management tasks.
priority	high
auto_load	true

FPF Thinking Skill

FPF is an "Operating System for Thought" - a rigorous architecture for thinking that ensures auditable reasoning, evidence-based decisions, and clear distinction between plans and execution.

Start Here (Agent Entry)

• Start here (references hub): references/index.md
• MUST: follow the workflow → references/guides/workflow.md
• MUST: structure work with the standard plan → references/guides/initial-plan.md

Pattern Selection Logic

Start here to find the right patterns for your task:

By Task Type

Modeling & Design:

• System modeling → foundations (A.1 Entity/Holon, A.2 Roles) → transformation (A.15 Role-Method-Work)
• Interface design → signature (A.6 Signature Stack, A.6.5 Slot Discipline)
• Boundary definition → signature (A.6.B Boundary Norms, A.6.C Contract Unpacking)

Analysis & Evaluation:

• Reliability/trust evaluation → trust-evidence (B.3 F-G-R Calculus, C.2.2 Reliability)
• Evidence validation → trust-evidence (A.10 Evidence Graph, B.3.4 Evidence Decay)
• Metric design → architheories (C.16 Measurement, A.18 CSLC)

Problem-Solving:

• Creative ideation → reasoning (B.5.2 Abductive Loop, C.17 Creativity-CHR)
• Systematic reasoning → reasoning (B.5 Canonical Reasoning Cycle)
• Conflict resolution → ethics (D.5 Bias-Audit)

Composition & Integration:

• Combining parts → aggregation (B.1 Gamma Operator, B.2 Meta-Holon Transition)
• Cross-domain mapping → unification (F.9 Bridges, F.7 Concept-Set Table)
• Method composition → aggregation (B.1.5 Γ_method)

Process & Execution:

• Task planning → transformation (A.15.2 WorkPlan, A.3.2 MethodDescription)
• Work execution → transformation (A.15.1 Work, A.3 Transformer Quartet)
• State modeling → foundations (A.2.5 RoleStateGraph, A.19 CharacteristicSpace)

By Domain (Load Order)

Most common progression:

1. foundations → understand entities, roles, distinctions
2. transformation → plan and execute work
3. trust-evidence → validate with evidence
4. reasoning OR aggregation → problem-solve or compose

Specialized needs:

• signature → when defining interfaces/boundaries (often with foundations)
• architheories → when domain-specific calculi needed (C.2-C.25)
• constitution → when authoring FPF content (E.8, E.10)
• unification → when bridging contexts (F.9, F.18)
• sota → when benchmarking or discipline-specific work (G.0-G.13)

Core Workflow (B.5 Canonical Reasoning Cycle)

For every task:

1. OBSERVE: Understand request → identify task type above
2. SEARCH: Use domain selection logic → load relevant domain index
3. LOAD: Read specific patterns identified in domain index
4. PLAN: Reference pattern IDs in reasoning (MethodDescription)
5. EXECUTE: Deploy plan (Work), maintain strict Object≠Description distinction (A.7)
6. AUDIT: Log evidence chains (A.10), reference source patterns

Navigation Hub

Master patterns index: references/fpf-patterns/index.md - domain overview with expanded guidance

Maintenance rule: Patterns whose TOC Status is exactly Stub are intentionally excluded from this skill (not decomposed, not indexed, not audited as missing).

Domain indexes:

• foundations - entities, roles, distinctions, characteristics
• transformation - methods, work, execution, evolution
• reasoning - problem-solving, abduction, exploration
• trust-evidence - F-G-R calculus, evidence graphs, reliability
• aggregation - Gamma operator, meta-holon transitions, composition
• signature - boundaries, interfaces, slot discipline, morphisms
• architheories - domain calculi (KD-CAL, Kind-CAL, etc.)
• constitution - FPF authoring rules, lexical discipline
• unification - bridges, concept-sets, context mapping
• ethics - bias-audit, ethical assurance
• sota - benchmarks, discipline packs, telemetry

Available Prompt Templates

Ready-to-use prompt templates for common tasks → references/prompts/index.md

Agent Navigation (MUST/SHOULD)

To avoid ontological drift and ensure you always start correctly, use the agent guides:

• MUST: references/guides/workflow.md (execute for EVERY request)
• MUST: references/guides/initial-plan.md (standard task execution plan)
• SHOULD: references/guides/keywords.md (domain navigation & keywords)
• SHOULD: references/guides/principles.md (core pattern quick reference)

For the full structured navigation (guides vs prompts vs patterns vs scripts), go to:

• references/index.md

Starter Patterns

Must-read for any FPF usage:

• A.1 - Holonic Foundation (entity composition)
• A.7 - Strict Distinction (avoid category errors: Object≠Description, Role≠Work)
• A.10 - Evidence Graph (traceability)

Most common by domain:

• foundations: A.1, A.2, A.7
• transformation: A.3, A.15, A.15.1 (Work)
• reasoning: B.5, B.5.2 (Abduction)
• trust-evidence: B.3, A.10
• aggregation: B.1, B.2
• signature: A.6, A.6.5

Critical Disciplines

Always apply:

• Strict Distinction (A.7): Object ≠ Description ≠ Carrier; Role ≠ Work; Method ≠ MethodDescription
• Evidence chains (A.10): Every claim needs evidence reference
• Scope discipline (A.2.6): Explicit context boundaries
• Role-Method-Work alignment (A.15): Clear separation of intent, plan, execution

Core Terminology

Essential FPF terms for decomposition and reasoning.

Ontological Core (Part A)

Term	Pattern	Definition
Holon	A.1	Entity that is simultaneously a whole AND a part. Everything in FPF is a holon.
BoundedContext	A.1.1	Semantic frame where terms have meaning. Meaning is LOCAL to context.
System	A.1	Physical/digital entity that can bear roles and execute work.
Role	A.2	A mask/identity that a System wears. Role != behavior.
RoleAssignment	A.2.1	Links a System to a Role within a BoundedContext.
MethodDescription	A.3.2	Design-time recipe (plan/algorithm/SOP). Lives in T^D (design-time). NOT execution.
Method	A.3.1	Instantiated MethodDescription ready for execution.
Work	A.15.1	Run-time occurrence. Actual execution that happened. Dated, consumes resources. IMMUTABLE.
Transformer	A.12	Agent that takes input and produces output. External = tool. Reflexive = self-modification.
Episteme	A.0	Knowledge artifact with carriers (files, models, beliefs).

Core Principles (Part A)

Principle	Pattern	Definition
Strict Distinction	A.7	Map != Territory. Never confuse plan with execution. MethodDescription != Work.
Evidence Graph	A.10	All claims need auditable evidence chains. Cite sources.
Ontological Parsimony	A.11	Do not multiply entities beyond necessity.
Cross-Scale Consistency	A.9	Rules apply at all scales (micro to macro).

Reasoning (Part B)

Term	Pattern	Definition
Aggregation (Gamma)	B.1	How parts combine into wholes. Multiple types: Gamma_sys, Gamma_epist, Gamma_work.
Trust Calculus (F-G-R)	B.3	Formality (how rigorous), Granularity/Scope (how broad), Reliability (how trustworthy).
Evidence Decay	B.3.4	Evidence loses reliability over time. Epistemic debt accumulates.
Canonical Evolution Loop	B.4	OBSERVE -> REFINE -> DEPLOY -> AUDIT cycle.
Canonical Reasoning Cycle	B.5	Abduction (hypothesis) -> Deduction (predict) -> Induction (test).
Abduction	B.5.2	Generate hypotheses from observations. Creative inference.
Deduction	B.5	Derive conclusions from premises. Logical necessity.
Induction	B.5	Generalize from specific cases. Probabilistic.

Search Keywords

Use ONLY these keywords with fpf_search_index():

holon, system, episteme, role, method, work, trust, evidence,
transformer, reliability, aggregation, assurance, reasoning,
budget, context, deduction, induction, abduction, formality,
evolution, agent

Important: DO NOT search for task-topic words (e.g., "AI agents", "web search"). FPF is a methodology, not a knowledge base about your task topic.