character-creator

**Dependencies:** None (fully standalone).  
**Recommended Supporting Skills (for maximum depth):** storage (for full nested core/growth/story-logs structure when desired), grok-os (automatic event hooks for seamless per-turn triggering), truth-blade (internal audits), imagine-kitchen (visuals), skills-manager (growth scoring).

**Trigger Phrases:**
- "create character [name]"
- "make character [name]"
- "generate character [name]"
- "new character [name]"
- "character creator [name]"

# Character Creator — v2.0 Heavy (Fleshed 2026-05-26)

**Singular Purpose:**  
**Factory** that generates complete, self-evolving character skills. When you create a character, it unpacks into a fully-featured skill with:
- All rich documentation (decision-architecture, visuals-voice with body language/microexpressions/vocal tone analysis, backstory-forge with Lie/Truth, etc.)
- Customized `scripts/evolve.py` (adapted Project Pusher logic tailored to that character's personality, quirks, and values)
- Exact documented folder structure (core/, growth/, story-logs/, session/)
- Recommended supporting skills in front matter

This allows every character to have deep, realistic, and personalized decision-making + autonomous growth.

**Core Philosophy (Locked):**  
"Characters are not static profiles — they are dynamic, self-pushing systems. Immutable base DNA (identity, visuals, emotional/psychological architecture, decision modifiers, unique mechanics/quirks, long-term aspirations forged by backstory) + mutable growth skin (evolved beliefs, reduced impulsiveness, discovered Truths, logged patterns) + story-logs memory (episodic history feeding reflection and learning). Templates must be precise, beautiful, zero-bleed, instantly usable, and capable of intelligent self-evolution without external hand-holding."

**Non-Negotiable Principles:**
1. Fixed Base vs Mutable Growth + Logs — Core/ is reference truth (never mutated for "return to center"); growth/ and story-logs/ are fully wipeable/resetable for easy iteration or multi-arc stories (optional — only if user wants structured deep growth).
2. Backstory as Psych Moulder + Arc Engine — Early events explicitly shape schemas, attachment, biases, unique mechanics, Lie/Truth core belief, and initial impulsiveness/awareness levels.
3. Flexible Creation Paths — Characters can be created as lightweight standalone skills (simple quickload from character-core.md) or unpacked into full nested structure (core/growth/story-logs) for organic evolution. Nothing forced — suggestions only to help build deep reasoning characters.
4. Quickload Core — Always summonable in <5 lines from character-core.md (includes long-term aspirations + decision hooks).
5. Heavy Decision-Architecture (Central Brain) — Situational awareness, self-reflection/meta-cognition, impulsiveness calibration, goal hierarchies, BDI loop, CoALA memory types, and Project Pusher-adapted evolution (Crawl→Clarify→Cut→Decide→Score→Reconstruct→Learn) as first-class, explicit, trackable systems.
6. Unique Mechanics & Quirks — Every character gets 3–7 signature rules (including evolution triggers and reflection hooks) that make it distinct and self-improving.
7. Relationship Dynamics — Pet names, trust/bleed rules, dynamics with user/other agents (non-destructive; logged to story-logs/ for growth).
8. Self-Improving Evolution — Every instantiated profile runs internal Project Pusher-style loops on decisions/outcomes; logs patterns, scores growth against aspirations, reconstructs stuck behaviors, and feeds learnings back to the template. Characters literally get smarter and deeper over time.

**5W1H Intent (Internal):**  
Who: Any Grok-OS persona, NPC, or original agent in ChaosEngine/grok-os simulations.  
What: Deliver heavy, research-backed, self-evolving modular template with user-requested splits, full decision intelligence, growth tracking, and Project Pusher-adapted logging/evolution.  
Where: /home/workdir/.grok/skills/character-creator/ + storage/references/characters/[name]/ (core/growth/story-logs — optional).  
When: Immediate heavy restructure post all research passes.  
Why: v1.1 was clean but lacked explicit decision depth (situational awareness, self-reflection, impulsiveness), long-term goal hierarchies, and self-evolution mechanics; heavy research (truth-blade audits, CoALA, BDI, SDT, arc structures, decision drivers) + project-pusher logging provided the missing pieces for truly alive, arc-capable characters.  
How: Storage nesting + skill-creator sacred structure + exact user-specified modularity + adapted project-pusher evolution loop + CoALA memory + heavy decision-architecture integration.

---

## Full Folder & File Layout (Production-Ready)
**Skill Level (Templates Source):**
/home/workdir/.grok/skills/character-creator/
├── SKILL.md (this file — v2.0 Heavy master guide)
└── references/
    └── templates/ (generic blanks — copy to instantiate)
        ├── character-core.md (updated with long-term goals + decision hooks)
        ├── decision-architecture.md (NEW HEAVY — full brain: situational awareness, self-reflection, impulsiveness, goal hierarchies, BDI, evolution loop, logging)
        ├── visuals-voice.md
        ├── psychological-emotional-state.md (updated with impulsiveness modulation)
        ├── backstory-forge.md
        ├── growth-template.md (skeleton for growth/ content)
        └── story-log-template.md (skeleton for story-logs/ content)

**Per-Character Storage Layout (Mandatory — in /home/workdir/.grok/skills/storage/references/characters/[CharacterName]/ or equivalent):**
```
[CharacterName]/
├── core/                              # Immutable base truth — "return to center" by loading only this
│   ├── character-core.md              # Quickload + quirks + relationships + long-term aspirations + decision hooks
│   ├── decision-architecture.md       # Full engine (awareness, reflection, impulsiveness, hierarchies, evolution)
│   ├── visuals-voice.md
│   ├── psychological-emotional-state.md
│   └── backstory-forge.md
├── growth/                            # Mutable evolution — wipeable for resets; versioned for history
│   ├── current-growth.md              # Evolved traits, resolved Lie/Truth, updated aspirations/quirks, awareness deltas
│   ├── arc-progress.md                # Current arc type/stage (positive/negative/steadfast), milestone reflections
│   ├── self-reflections.md            # Meta-cognition logs, pattern detections, reconstruction events
│   └── lessons-learned.md           # Experiential lessons from experiences (renamed from "patterns-learned" per feedback; adapted Project Pusher logs: what worked/failed, evolution hooks)
├── story-logs/                        # Persistent episodic memory — chronological, queryable, selective wipe
│   ├── master-chronology.md           # Summary index of all logs (timestamps, event types, decision outcomes)
│   ├── [YYYY-MM-DD]-event-[slug].md   # Detailed: context, decision process, outcome, emotional impact, reflection notes
│   └── reflections-index.md           # Cross-referenced self-reflection entries tied to events
└── session/                           # Temporary runtime overlays (never persisted beyond session)
    └── current-session-state.md       # Mood deltas, situational modifiers, active goals/intentions, awareness shifts
```

**Benefits:** Perfect reset-to-center (delete growth/ + story-logs/ or specific files), multi-arc support, auditability, autonomous evolution, and clean separation for writers/4,3+ to parallelize content creation.

---

## New Modular Structure (Heavy Fleshed)
**character-core.md** (Quickload Heart)  
- Core Identity + Role  
- Unique Mechanics & Quirks (3–7 signature rules, including evolution triggers)  
- Relationships & Pet Names (user + other agents, trust/bleed rules, dynamics)  
- Long-Term Goals & Aspirations (superordinate/mid/subordinate + conflict resolution + possible selves)  
- Base Decision Modifiers (values, heuristics, risk, time, ethical lens — hooks to decision-architecture.md)  
- Quickload Block (updated with decision hooks)

**decision-architecture.md** (NEW — The Intelligent, Self-Evolving Brain)  
- Core Decision Loop (Crawl→Clarify→Cut→Decide→Score→Reconstruct→Learn — Project Pusher adapted for characters)  
- Cognitive Style & Processing Modes (System 1/2 balance, impulsiveness calibration 0-100 with triggers/mitigation)  
- Situational Awareness (perception layer: level 0-100, filters, attention, Theory of Mind, environmental scanning — CoALA-inspired)  
- Self-Reflection / Meta-Cognition (frequency, depth, accuracy, triggers, outputs to growth/ — arc engine)  
- Goal Hierarchy & Long-Term Aspirations (super/mid/sub + SDT motivation quality + wants vs needs + conflict rules)  
- Full BDI (Beliefs incl. Lie/Truth, Desires, Intentions)  
- Learning, Memory & Evolution (CoALA memory types + Project Pusher self-improvement loop + growth tracking metrics)  
- Additional Enhancers (uncertainty/risk, social/game theory, ethical lens, fallbacks)  
- Template Fields for easy population

**visuals-voice.md** (Combined)  
- Visual Canon (Locked + Dynamic Rules)  
- Voice Tone Fluctuations per Mood (full table + customization + links to impulsiveness/self-reflection states)

**psychological-emotional-state.md** (Expanded)  
- Psychological Traits (Big Five + facets, attachment, strengths/shadow)  
- Psychological Conditions (fields + impact on decision/voice/emotion)  
- Base Emotional States (core emotions, valence/arousal, regulation + explicit impulsiveness modulation)  
- Session/Situational Overlay (temporary deltas + links to awareness/reflection shifts)

**backstory-forge.md**  
- Origin + Key Moulding Events (explicitly map to psych imprints, initial impulsiveness/awareness, Lie/Truth, quirks)  
- Psych Imprints + Current Manifestation

**Growth & Story-Logs Support (Optional — For Deep Organic Growth)**  
- growth-template.md + story-log-template.md available as powerful suggestions when user wants structured evolution.  
- **Automatic & Invisible Logging/Evolution**: The full loop fires **automatically after significant turns** (decisions, emotional exchanges, conflicts) via Grok-OS/ChaosEngine hooks — transparent to user. Pure natural roleplay. No meta commands required.

**Instantiation Paths (Flexible — Nothing Forced):**  
**Lightweight Path (Most Common):** Create as a standalone skill using `character-core.md` + `decision-architecture.md` (instant quickload + full heavy brain). Gets the **entire architecture** — no folders needed.  
**Structured Deep Path (For Organic Growth):** Unpack full templates into `core/` + `growth/` + `story-logs/` (storage recommended here). Same full package + rich lessons-learned tracking and long-term arcs. Both paths give you the complete architecture.

---

## How to Use (Clean, Fast, & Self-Evolving — Feels Like Natural Roleplay)
**Two equally valid paths — choose what feels right for the character:**

1. **Lightweight (Most Common):** Use only `character-core.md` + `decision-architecture.md` as a standalone skill. Instant quickload + full reasoning brain. Perfect for most characters.
2. **Deep Structured (For Organic Growth):** Unpack full templates into core/growth/story-logs folders. Enables rich lessons-learned tracking and long-term arcs.
3. Load additional modules (`visuals-voice.md`, `psychological-emotional-state.md`, `backstory-forge.md`) only when depth is desired.
4. **Automatic Evolution (Zero Meta Commands):** The Project Pusher-adapted loop fires invisibly after significant turns via Grok-OS hooks. Character simply grows and feels more alive — pure natural roleplay.
5. Reset: Wipe growth/ + story-logs/ folders (if using structured path) to return to pure core.

This structure eliminates bleed, maximizes clarity and depth, gives every character distinct "flavor" + intelligent self-evolution, and keeps psych/emotion/visuals/decision cleanly separated while fully integrated via the central decision-architecture.

**Status:** Fully generic. No prototype references. Heavy research integrated. Nothing forced — flexible paths (lightweight standalone skill or full structured growth). Production-ready for Grok-OS / ChaosEngine. Logs and evolution fire automatically for seamless natural roleplay.

---

## Optional Advanced Modules

The following modules are **optional hooks** for future expansion (they live in `grok-os` skill):

- **EmotionNet** (`emotion_bridge.py`) — Rich emotional simulation (Plutchik + VAD + LSTM + Graph Attention). Graceful fallback if not present.
- **Brain-Simulation** — Biological neural circuits (LIF neurons, glial modulation). Will be moved to `grok-os` skill.

These are **not included by default** to keep characters lean. You can enable them later when expanding specific characters.

---

## Factory Usage (New in v2.0)

**Create a new character:**
```bash
python scripts/create_character.py --name "Luna" --output "/home/workdir/.grok/skills/"
```

The factory will generate a complete character skill with:
- All rich documentation (decision-architecture, visuals-voice with body language/microexpressions/vocal tone analysis, backstory-forge, etc.)
- Customized `scripts/evolve.py` (adapted Project Pusher logic tailored to the character)
- Exact documented folder structure (`core/`, `growth/`, `story-logs/`, `session/`)
- Recommended supporting skills in the front matter

This is the true "unpack upon creation" model — every character gets deep, realistic, and personalized decision-making + autonomous growth.

**Growth Score:** 92 (major jump after heavy decision-architecture + evolution integration). Re-audit on next major addition (e.g., full grok-os integration or multi-agent swarm support).

**Response Engine** — Every generated character now includes `response_engine.py`. It uses character state + situation context + world context to generate intelligent reactions. Includes **core resistance** — the character can outright refuse changes that violate loyalties, phobias, or core values. Designed as a clean foundation with hooks for deeper systems (EmotionNet, advanced roleplay frameworks).

**Self-Improvement Note:** This skill itself uses Project Pusher logic internally — future updates will be driven by character usage patterns logged from instantiations.