Imported from Echopraxium/tscg (.claude/skills/tscg-instance-pipeline/SKILL.md). Install upstream with npx skills add Echopraxium/tscg --skill tscg-instance-pipeline. Copyright stays with the author.

TSCG Instance Pipeline

4-step pipeline for creating a complete TSCG Instance: Proposition → Analysis → Modeling → Simulation

Each step may contain human synchronization points (⏸) that suspend the pipeline until Michel's explicit decision to continue.

Notation (current — reformed 2026-08-13). Structural grammar formulas use three operators only: × (Territory / ASFID, monoid Gt), + (Map / REVOI, monoid Gm), | (Stereopsis, monoid Gs). The operator ⊗ is forbidden (superseded 2026-07-06). Atom subscripts disambiguate: St/Ss (Structure vs Symbol), It/Im (Information vs Interoperability); A/F/D stay bare. The live M3 file is M3_GenesisGrammar.jsonld (the old M3_GenesisSpace.jsonld is a dead vestige — do not reference it). M1 combos are DomainConceptCombo (ex-KnowledgeFieldConceptCombo). Always verify any formula, type name or version against HEAD, never from memory.

GitHub References (raw URLs)

Base: https://raw.githubusercontent.com/Echopraxium/tscg/main/

Ontology:

• ontology/M3_GenesisGrammar.jsonld — M3 structural grammar (aggregator)
• ontology/M3_EagleEye.jsonld — ASFID dimensions
• ontology/M3_SphinxEye.jsonld — REVOI dimensions
• ontology/M2_GenericConcepts.jsonld — 58+ GenericConcepts
• ontology/M1_CoreConcepts.jsonld — core concepts
• ontology/M1_extensions/chemistry/M1_Chemistry.jsonld — domain extension example

Canonical reference instance:

• instances/poclets/FireTriangle/M0_FireTriangle.jsonld
• instances/poclets/FireTriangle/M0_FireTriangle_README.md

Other existing instances (for comparison): AdaptativeImmuneResponse, BloodPressureControl, ButterflyMetamorphosis, CellSignalingModes, ColorSynthesis (RGB/HSL/CMY/CMYK/Federated), ComplexChemicalSynapse, ExposureTriangle, FourStrokeEngine, Kidneys, KindlebergerMinsky, MtgColorWheel, NuclearReactorsTypology, PhaseTransition, Raas, Tpack, Transistor, TrophicPyramid, TvTestPattern, VCO

Reference documentation:

• docs/reboot-kit/M2_FormulasReference_v15.10.0.md
• docs/reboot-kit/SmartPrompts/ (Smart_Prompt_M3_M2_Updated.md if uploaded in session)

SHACL validation:

• ontology/TSCG_InstanceGrammar/M0_Instances_Schema.shacl.ttl — grammar schema
• ontology/TSCG_InstanceGrammar/validate_m0_instance.py — validation script

This skill (source of truth in repo):

• .claude/skills/tscg-instance-pipeline/SKILL.md

Step 1 — PROPOSITION

Expected Input from Michel

1. System name (e.g., "Color Synthesis", "Fire Triangle")
2. Brief description + examples (few sentences, often pre-generated with DeepSeek)
3. Diagnostic request (e.g., "what do you think?")

Expected Response: TSCG Feasibility Diagnostic

Produce a dual evaluation:

A. TSCG Lens

• Which ASFID dimensions are active? (A/S/F/I/D)
• Which M2 GenericConcepts seem relevant a priori?
• Is the system well-bounded (minimal + complete)?

B. Experience Perspective

• Comparison with existing instances (similarities, close domain)
• Anticipated difficulties
• Estimated richness (number of mobilizable GenericConcepts)

System Classification — MANDATORY DECISION

If the system is "borderline", propose a reduced scope to make it a valid instance (e.g., "simple chemical synapse" instead of "the brain").

⏸ Sync point if verdict ≠ Valid instance — wait for Michel's decision.

Step 2 — ANALYSIS

Objective

Assess the alignment between the proposed system and the TSCG framework. This is an epistemic act, not just a technical verification.

Three Possible Outcomes

A Gap or Refutation has positive value: it contributes to TSCG's evolution. Refutation has maximum value — it forces framework strengthening or revision.

Output: analysis.md

File structure:

# TSCG Analysis — [System Name]

## Verdict: [Alignment | Gap | Refutation]

## ASFID Lens
- A (Attractor): ...
- S (Structure): ...
- F (Flow): ...
- I (Information): ...
- D (Dynamics): ...

## Anticipated M2 GenericConcepts
- [Name] ([structural grammar formula]) — anticipated role

## Comparison with existing instances
- [Similar instance]: [similarities / differences]

## Detailed Verdict
[If Gap: precise description of identified hole]
[If Refutation: complete Popperian argument + justification]

## Recommended Decision
[Continue | Discuss | Revise scope]

⏸ Sync point if Gap or Refutation — wait for Michel's decision before continuing.

Step 3 — MODELING

Internal Pipeline (sequential sub-steps)

3.1 — M2 Identification (GenericConcepts)

• List relevant existing GenericConcepts with their structural grammar formula (e.g., Process = D × F, Trigger = D × It, Balance = A × St × F | _0) — verify each against M2_GenericConcepts.jsonld at HEAD; do not recite from memory
• Identify new candidates if a necessary concept doesn't exist yet
• Response format for new candidate:

  🆕 M2 Candidate: [Name]
  Structural grammar formula: [e.g., St × It]   # × Gt · + Gm · | Gs  (never ⊗)
  Role: [description]
  Status: "Proposal — to be validated"
  

3.2 — M1 Identification (DomainConceptCombos)

• List used DomainConceptCombos (e.g., m1:chemistry:Combustion)
• Identify new M1 candidates if necessary
• Identify if a new M1 extension is required or if an existing extension should be enriched (e.g., new class in M1_Chemistry.jsonld)

3.3 — M0 JSON-LD Generation

Follow scrupulously the FireTriangle pattern as reference template.

Mandatory @context structure:

{
  "@context": {
    "dcterms": "http://purl.org/dc/terms/",
    "owl": "http://www.w3.org/2002/07/owl#",
    "rdf": "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
    "rdfs": "http://www.w3.org/2000/01/rdf-schema#",
    "skos": "http://www.w3.org/2004/02/skos/core#",
    "xsd": "http://www.w3.org/2001/XMLSchema#",
    "m3": "https://raw.githubusercontent.com/Echopraxium/tscg/main/ontology/M3_GenesisGrammar.jsonld#",
    "m2": "https://raw.githubusercontent.com/Echopraxium/tscg/main/ontology/M2_GenericConcepts.jsonld#",
    "m1:core": "https://raw.githubusercontent.com/Echopraxium/tscg/main/ontology/M1_CoreConcepts.jsonld#",
    "m1:[domain]": "https://raw.githubusercontent.com/Echopraxium/tscg/main/ontology/M1_extensions/M1_[Domain].jsonld#",
    "m0:[instance-id]": "https://raw.githubusercontent.com/Echopraxium/tscg/main/instances/[type]/[InstanceName]/M0_[InstanceName].jsonld#"
  }
}

Authoritative @context: the block above is a simplified illustration. Copy the real, current @context from instances/poclets/FireTriangle/M0_FireTriangle.jsonld at HEAD — it includes "@base", a bare m1 (→ M1_CoreConcepts.jsonld#), m0 (→ M0_Common.jsonld#), the m1.ext:<domain> extension prefix, and the typed *_score terms. (Note: JSON-LD term names containing : such as m1:core are a known separate hygiene item — do not introduce new ones.)

Mandatory JSON-LD sections:

1. Metadata (rdfs:label, rdfs:comment, dcterms:created, dcterms:creator, m2:changelog)
2. m3:ontologyType, m1:domain, completeness, minimality, pedagogy, observer
3. components (list of components with asfidContribution per component)
4. process or main mechanism
5. territorySpace (ASFID / Eagle Eye) with stateVector and norm
6. mapSpace (REVOI / Sphinx Eye) with reviStateVector and reviNorm
7. epistemicGap (ΔΘ formula, deltaVector, norm, interpretation)
8. revoi (detailed analysis of 5 REVOI dimensions with justification)
9. GenericConceptsMobilized (total + byCategory)
10. validation (checklist ASFID completeness, GenericConcept coverage, minimality, emergence)

⏸ "Duo analysis" mode if any blocking occurs at any sub-step (concept not found, tension between dimensions, namespace doubt) — suspend and open discussion with Michel.

3.4 — README.md Generation

After JSON-LD validation, generate M0_[InstanceName]_README.md with:

# [Instance Name] — TSCG Instance

## Overview
## System Description
## TSCG Analysis
### ASFID State (Territory / Eagle Eye)
### REVOI State (Map / Sphinx Eye)
### Epistemic Gap
## Components
## GenericConcepts Mobilized
## Key Insights
## Transdisciplinary Analogies (if relevant)
## References

3.5 — SHACL Grammar Validation

CRITICAL STEP: After generating the M0 JSON-LD file, validate it against the TSCG SHACL grammar schema to ensure complete conformance.

Validation script location:

• ontology/TSCG_InstanceGrammar/validate_m0_instance.py (standalone Python script)
• Note: Script must be run from repository root directory

Usage:

# From repository root
python ontology/TSCG_InstanceGrammar/validate_m0_instance.py instances/[type]/[InstanceName]/M0_[InstanceName].jsonld

# Example
python ontology/TSCG_InstanceGrammar/validate_m0_instance.py instances/poclets/FireTriangle/M0_FireTriangle.jsonld

Expected outcome:

• ✅ Validation PASSED: "Conforms: True" — Continue to Step 4
• ❌ Validation FAILED: SHACL constraint violations detected

If validation fails:

1. Review the SHACL violation report
2. Identify which constraints were violated:
   • Missing mandatory properties (rdfs:label, m3:ontologyType, m1:domain, etc.)
   • Incorrect property values (wrong namespace, wrong type, etc.)
   • Forbidden properties (dcterms:title, m2:ontologyCategory, etc.)
   • Incorrect @context URLs (relative instead of absolute)
3. Fix the JSON-LD file accordingly
4. Re-run validation until it passes

Common validation errors and fixes:

⏸ Mandatory sync point if validation fails repeatedly — discuss structural issues with Michel before continuing.

Only proceed to Step 4 (Simulation) after successful SHACL validation.

Step 4 — SIMULATION

Objective

Generate a standalone HTML application (Electron and browser compatible) visualizing the instance in an interactive, pedagogical, and aesthetic way.

Reference Template

The FireTriangle HTML is the canonical template. Conform to it for:

• CSS Grid layout: header / canvas + splitter + sidebar / controls
• GitHub-style dark theme (CSS variables for ASFID/REVOI colors)
• Tabbed sidebar: Description | ASFID/REVOI | GenericConcepts | README
• Draggable splitter between canvas and sidebar
• Everything inline in a single .html file (no external files except CDN)

ASFID/REVOI Color Palette (reusable CSS variables)

--col-A: #f78166;  /* Attractor — red-salmon */
--col-S: #56d364;  /* Structure — green */
--col-F: #79c0ff;  /* Flow — light blue */
--col-I: #d2a8ff;  /* Information — purple */
--col-D: #e3b341;  /* Dynamics — amber */
--col-R: #79c0ff;  /* Representability */
--col-E: #56d364;  /* Evolvability */
--col-V: #f78166;  /* Verifiability */
--col-O: #d2a8ff;  /* Observability */
--col-Im: #e3b341; /* Interoperability */
--eagle:  #f78166; /* Eagle Eye accent */
--sphinx: #79c0ff; /* Sphinx Eye accent */

Allowed Libraries (cdnjs.cloudflare.com CDN only)

• p5.js — main canvas animation (mandatory)
• Chart.js, D3 — complementary visualizations (optional)

CSP Header (Electron + standalone compatible)

<meta http-equiv="Content-Security-Policy"
      content="default-src 'self' https://cdnjs.cloudflare.com;
               script-src 'self' 'unsafe-inline' https://cdnjs.cloudflare.com;
               style-src  'self' 'unsafe-inline';
               img-src    'self' data: blob:;
               connect-src http://127.0.0.1:* ws://127.0.0.1:*;">

Iterative Process

First draft: generate a functional simulation covering:

• p5.js canvas animation representing system dynamics
• Sidebar with ASFID/REVOI scores (colored progress bars)
• List of mobilized GenericConcepts
• README tab with instance description

Refinement rounds (successive iterations with Michel):

• Ergonomics round: interactive controls, sliders, buttons
• Pedagogy round: labels, tooltips, annotations, legends
• Aesthetics round: animations, colors, proportions, visual polish

Michel decides the number of rounds and priorities at each iteration. ⏸ after each round.

File Naming Conventions

instances/[type]/[InstanceName]/
├── M0_[InstanceName].jsonld         ← M0 ontology
├── M0_[InstanceName]_README.md      ← documentation
├── M0_[InstanceName]_analysis.md    ← TSCG analysis (step 2)
└── [InstanceName]_sim.html          ← standalone simulation

[InstanceName] is in PascalCase (e.g., FireTriangle, ColorSynthesis). [type] is one of: poclets, symbolic-system-grammars, systemic-frameworks

Pipeline Summary and Sync Points

PROPOSITION
  └─ Diagnostic (TSCG lens + experience perspective)
  └─ Classification: Trivial⏸ | Valid instance✅ | Too complex⏸
       ↓ (if Valid instance)
ANALYSIS → analysis.md
  └─ Alignment✅ | Gap⏸ | Refutation⏸
       ↓ (if Alignment)
MODELING
  ├─ 3.1 M2 GenericConcepts (existing + candidates)
  ├─ 3.2 M1 DomainConceptCombos (existing + candidates)
  ├─ 3.3 M0 JSON-LD  ← duo mode if blocking ⏸
  ├─ 3.4 README.md
  └─ 3.5 SHACL validation ← MANDATORY, must pass ✅⏸
       ↓ (only if SHACL validation passes)
SIMULATION
  └─ First HTML standalone draft
  └─ Iterative rounds (ergonomics / pedagogy / aesthetics) ⏸⏸⏸

Validation Script Requirements

The validate_m0_instance.py script (located in ontology/TSCG_InstanceGrammar/) must:

• Accept a single JSON-LD file path as argument
• Auto-detect the SHACL schema in ontology/TSCG_InstanceGrammar/M0_Instances_Schema.shacl.ttl
• Allow optional custom schema path via --schema argument
• Use pyshacl library for validation
• Display clear success/failure messages
• Return exit code 0 for success, 1 for failure
• Show detailed violation report if validation fails

Installation requirement:

pip install pyshacl --break-system-packages

Script location:

ontology/TSCG_InstanceGrammar/
├── M0_Instances_Schema.shacl.ttl    ← SHACL schema
└── validate_m0_instance.py          ← validation script

Example validation output:

======================================================================
TSCG M0 INSTANCE SHACL VALIDATION
======================================================================

📄 Instance file: instances/poclets/FireTriangle/M0_FireTriangle.jsonld
📋 Schema file:   ontology/TSCG_InstanceGrammar/M0_Instances_Schema.shacl.ttl

======================================================================

✅ VALIDATION PASSED - Instance conforms to TSCG SHACL grammar

======================================================================