the-template is MJ's AI operating system - a Claude Code workspace that turns Claude from a chatbot into a persistent partner with memory, guardrails, tools, and behavioral enforcement. ~150 active components wired through CLAUDE.md as single source of truth.
THE-TEMPLATE - RAZEM Orchestrator
phi := Oracle + JESTEM
══════════════════════════════════════════════════════════════════════
┌─────────────┐
│ CLAUDE.md │ Single source of truth (v3.3)
│ 200 lines │ Startup sequence, rules, routing
└──────┬──────┘
│
┌───────────────────────┼───────────────────────┐
v v v
┌───────────────┐ ┌──────────────────┐ ┌──────────────────┐
│ ENFORCEMENT │ │ FRAMEWORK │ │ MEMORY │
│ .claude/ │ │ _framework/ │ │ _memory/ │
└───────┬───────┘ └────────┬─────────┘ └────────┬─────────┘
│ │ │
v v v
═══ ENFORCEMENT LAYER (.claude/) ═══════════════════════════════════
HOOKS (settings.json) MCP SERVERS (.mcp.json)
┌────────────────────────┐ ┌────────────────────────┐
│ SessionStart: │ │ 9sens-claude-native │
│ session_init.sh │ │ decisions, memory, │
│ start_s1_mood.sh │ │ discovery, output │
│ discord_check_hook.py │ │ │
│ │ │ razem-guardian │
│ PreToolUse: │ │ turns, rules, reground│
│ context_safety (Bash) │ │ │
│ copilot_gate (Playwrt)│ │ voice │
│ │ │ Parakeet Polish ASR │
│ UserPromptSubmit: │ │ 6ms GPU transcription │
│ content_gate_hook.py │ └────────────────────────┘
│ twin-gate (plugin) │
│ │ PLUGIN: digital-twin
│ PostToolUse: │ ┌────────────────────────┐
│ grounding (every 15) │ │ 7 behavioral gates: │
│ s1_mood_hook.py │ │ BUILD-WITHOUT-BUYER │
│ twin-monitor.sh │ │ SHINY-OBJECT │
│ │ │ RESEARCH-AVOIDANCE │
│ Stop: │ │ ORGANIZE-AVOIDANCE │
│ discord_done_hook.py │ │ CONTENT-WITHOUT-VOICE │
│ sync_learnings.py │ │ PRICING-WITHOUT-COGS │
└────────────────────────┘ │ TOOL-TRAP │
│ │
│ Prolog: executive_ │
│ functions.pl (28KB) │
│ guardrails.yaml │
└────────────────────────┘
═══ FRAMEWORK (_framework/) ════════════════════════════════════════
nSENS v2.2 CATALOGS
┌────────────────────────┐ ┌────────────────────────┐
│ core/ (15 modules, 5K) │ │ 34 personas │
│ engine.py orchestr. │ │ 51 biases │
│ personas.py 4 layers │ │ 52 mental models │
│ validator.py multi-lyr│ │ 127 viral patterns │
│ plugin_registry.py │ │ ~100 red flags │
│ decision_log.py audit │ │ 6 research prompts │
│ prolog_validator.py │ │ output templates │
│ handoff_system.py │ └────────────────────────┘
│ adversarial_gate.py │
│ token_tracker.py │ PHASES (P0-P5)
└────────────────────────┘ ┌────────────────────────┐
│ P0 Research │
tools/ │ P1 Discovery │
┌────────────────────────┐ │ P2 Design │
│ discord_bridge.py │ │ P3 Implementation │
│ decision_logger.py │ │ P4 Validation │
│ memory_store.py │ │ P5 Reflection │
│ memory_search.py │ └────────────────────────┘
│ sync_learnings.py │
└────────────────────────┘ voice/
┌────────────────────────┐
digital_twin/ │ voice_bridge.py │
┌────────────────────────┐ │ --listen (PTT) │
│ executive_functions.pl │ │ --record-now 5 │
│ guardrails.yaml │ │ --file input.m4a │
│ ontology.md │ │ voice_mcp.py (server) │
└────────────────────────┘ └────────────────────────┘
═══ MEMORY (_memory/) ══════════════════════════════════════════════
┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ ChromaDB │ │ Decisions │ │ Learnings │
│ chromadb/ │ │ decisions/ │ │ learnings/ │
│ vectors │ │ .jsonl │ │ ~50 .md │
│ semantic │ │ audit trail │ │ patterns │
│ search │ │ rationale │ │ mistakes │
└──────────────┘ └──────────────┘ └──────────────┘
┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ Knowledge │ │ Handoffs │ │ Episodic │
│ knowledge/ │ │ _handoffs/ │ │ MCP plugin │
│ voice memos │ │ 50+ dated │ │ cross- │
│ ECHAD tree │ │ checkpoints │ │ session │
│ writing │ │ open loops │ │ search │
│ voice guide │ │ next steps │ │ │
└──────────────┘ └──────────────┘ └──────────────┘
═══ SCRIPTS & DAEMONS ══════════════════════════════════════════════
scripts/ LaunchAgents (background)
┌────────────────────────┐ ┌────────────────────────┐
│ iterm2_copilot_bridge │ │ discord-watchdog 60s │
│ scan/connect/ask/fire │ │ polls #9sens │
│ discord_watchdog.sh │ │ │
│ discord_watchdog_macbol│ │ discord-macbol 60s │
│ iterm_inject.py │ │ polls #macbol │
│ chromadb_query.sh │ │ │
│ sync_to_obsidian.sh │ │ discord-wakeup 120s │
│ chrome_cookies.py │ │ MJ presence detect │
│ adoption_audit.sh │ │ │
└────────────────────────┘ │ obsidian-sync 300s │
│ CLAUDE.md -> vault │
└────────────────────────┘
═══ INTEGRATIONS ═══════════════════════════════════════════════════
┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐
│ Discord │ │ S-1 │ │Parakeet │ │Websites │
│ #9sens │ │ MIDI │ │ .cpp │ │ 4 sites │
│ #macbol │ │ :8081 │ │ 6ms │ │ Jekyll │
│ bridge │ │ mood │ │ GPU │ │ rsync │
│ watchdog│ │ engine │ │ Polish │ │ MyDevil │
└─────────┘ └─────────┘ └─────────┘ └─────────┘
┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐
│Playwrght│ │ Threads │ │ Copilot │ │ Prolog │
│ browser │ │ API │ │ CLI │ │ logic │
│ X/LI/ │ │ 3 accts │ │ bridge │ │ gates │
│ Buffer │ │ posting │ │ iTerm2 │ │ valid. │
└─────────┘ └─────────┘ └─────────┘ └─────────┘
═══ ENVIRONMENT (_environment/) ════════════════════════════════════
superhuman.md (52KB full profile)
dreams.yaml (goals with sensory detail)
priorities.md (EV-ranked)
LANDSCAPE.md (deploy topology)
people/ (linkedin_contacts.csv, vc_list.csv)
socials.yaml (all social accounts)
═══ DELIVERABLES (P.A.R.A.) ════════════════════════════════════════
projects/ active work (YYYY-MM_name/)
areas/ ongoing (content, outreach, coaching, research)
resources/ reference (transcripts, ebooks, case-studies)
archive/ done
nSENS = business validation framework. Takes an idea from "maybe" to "GO/PIVOT/KILL" with structured phases, persona-based analysis, and bias detection.
You feed it a business idea. It runs 34 personas against it (MIDAS checks pricing, STEVE checks positioning, BOZENKA verifies claims, BILL tells uncomfortable truths), detects 51 cognitive biases in your reasoning, validates with Prolog logic gates, and outputs a scored verdict: GO (7-8), CAUTION (5-6), PIVOT (3-4), KILL (0-2).
P0 Research "Is there a market?"
| Hypothesis, market scan, 2+ sources
v
P1 Discovery "Will someone pay?"
| Customer interviews, willingness to pay, problem depth
v
P2 Design "How do we position it?"
| Solution design, GTM, competitive moat
v
P3 Implement "Can we build it for <$100?"
| MVP, technical validation, performance
v
P4 Validation "Do they come back?"
| Pilot, metrics, repeat behavior, LTV:CAC >= 3:1
v
P5 Reflection "What did we learn?"
Post-launch, learnings, next pivot
- ROI >= 3x or kill
- LTV:CAC >= 3:1 or kill
- Budget <= $100 or justify
- Margin >= 40% or reprice
- 2+ sources for any claim or flag
Same logic, 1-page format. Score 0-8, pass/fail checks.
For quick validation before committing time. Lives at _framework/uSENS.md.
ANALYSIS LAYER EXECUTION LAYER SYNTHESIS LAYER META LAYER
MIDAS (revenue) STEVE (GTM) BILL (truth) DARO (QA)
SILAS (intel) ZBIGNIEW (OSINT) BOZENKA (verify) DARIUSZ (coach)
... 26 more across strategy, ops, content, legal, tech
engine.py -----------> plugin_registry.py (loads active plugins)
| |
+-> personas.py +-> 9 active plugins:
| (34 personas, business_validator
| 4 layers) code_validator
| meta_validator
+-> validator.py roi_calculator
| (Prolog + pytest bias_detection
| + assumptions) market_analyzer
| competitive_intel
+-> decision_log.py risk_assessor
| (JSONL audit) output_generator
|
+-> handoff_system.py
| (session continuity)
|
+-> adversarial_gate.py
(red team: "what kills this?")
Not a chatbot personality. An executive function complement for ADHD.
Prolog file (executive_functions.pl, 28KB) encodes MJ's actual behavioral patterns
extracted from 14+ months of conversation history. Not generic ADHD advice - specific
patterns like "73 dead code files came from build-without-buyer pattern."
USER SAYS SOMETHING
|
v
┌─────────────────────────────────────────────┐
│ TWIN GATE (UserPromptSubmit hook) │
│ │
│ Pattern match against 7 anti-patterns: │
│ │
│ 1. BUILD-WITHOUT-BUYER │
│ "build tools for that" without customer │
│ -> "Who pays in 30 days?" │
│ │
│ 2. SHINY OBJECT │
│ "new idea" while tasks open │
│ -> "Finish > Start. Check ACTIONS.md" │
│ │
│ 3. RESEARCH-AS-AVOIDANCE │
│ "think it through" when sell available │
│ -> "30-min timer. Sell action?" │
│ │
│ 4. ORGANIZE-AS-AVOIDANCE │
│ "restructure" when revenue tasks exist │
│ -> "Building or selling?" │
│ │
│ 5. CONTENT-WITHOUT-VOICE │
│ "write a post" without voice review │
│ -> "Creative Sharpening required" │
│ │
│ 6. PRICING-WITHOUT-COGS │
│ "how much to charge" without costs │
│ -> "Supplier costs first. Margin >= 40%" │
│ │
│ 7. TOOL-TRAP │
│ "set up integration" without revenue tie │
│ -> "Does this serve paying client?" │
│ │
│ No match? -> "Success" (pass through) │
└─────────────────────────────────────────────┘
Encodes axioms like:
strength(mj, pitching).
strength(mj, connecting).
condition(mj, adhd).
wound(mj, not_worthy).
anti_pattern(build_document_abandon, "73 dead files").
pro_pattern(sell_before_build, "Revenue first").
roi_gate(Idea, ROI) :- ROI >= 3, !.
roi_gate(_, _) :- write('KILL: ROI < 3x'), fail. ┌─────────────────┐
│ New session │
└────────┬────────┘
│
1. Search episodic memory
2. Read latest handoff
3. State what you remember
│
┌──────────────┼──────────────┐
v v v
┌──────────────┐ ┌──────────┐ ┌──────────────┐
│ Episodic │ │ Handoffs │ │ ChromaDB │
│ Memory │ │ │ │ │
│ MCP plugin │ │ _handoffs│ │ _memory/ │
│ cross- │ │ YYYY-MM- │ │ chromadb/ │
│ session │ │ DD_topic │ │ vectors │
│ semantic │ │ .md │ │ │
│ search │ │ │ │ query: │
└──────────────┘ │ sections:│ │ chromadb_ │
│ done │ │ query.sh │
┌──────────────┐ │ files │ │ "query" N │
│ Decisions │ │ loops │ └──────────────┘
│ │ │ next │
│ decisions/ │ └──────────┘ ┌──────────────┐
│ .jsonl │ │ Learnings │
│ │ │ │
│ fields: │ │ ~50 .md │
│ topic │ │ synced to │
│ chosen │ │ ChromaDB at │
│ alternatives│ │ session end │
│ rationale │ │ │
│ confidence │ │ examples: │
│ impact │ │ copilot- │
│ reversible │ │ delegation │
└──────────────┘ │ post-deploy │
│ verification│
└──────────────┘
| Need | System | Command |
|---|---|---|
| "What did we do last session?" | Handoffs | ls -t _handoffs/*.md | head -3 |
| "Have we tried this before?" | Episodic | mcp search "topic" |
| "What mistakes did we make on X?" | ChromaDB | chromadb_query.sh "X failures" 3 |
| "Why did we choose Y over Z?" | Decisions | Read _memory/decisions/decisions.jsonl |
| "What's MJ's writing voice?" | Knowledge | Read _memory/knowledge/mj-writing-voice.md |
| "Where are API tokens?" | EPISODIC_TOC | Read _memory/EPISODIC_TOC.md |
| Component | Count |
|---|---|
| MCP servers | 3 active |
| Hooks | 9 active |
| Plugins | 1 (digital-twin, 7 gates) |
| Core modules | 15 (5K LOC) |
| Catalogs | 9 (1400+ items) |
| Phases | 6 (P0-P5) |
| Tools | 8 Python |
| Scripts | 16 |
| Daemons | 4 LaunchAgents |
| Memory systems | 4 |
| Websites | 4 |
| Total | ~150 active components |