| name | corey-mcclain-persona-engineering |
| description | Adds a narrative identity layer (Persona) on top of Logic, Library, and Memory to produce AI agent outputs that cross from "technically accurate" to "recognizably distinct." The LLMP framework (Logic → Library → Memory → Persona) plus the Makoshi state-capture protocol transforms how agents think by changing the world they inhabit. Use when designing AI agents that need a distinct voice / character, when generic agent outputs feel interchangeable, when building expert personas with embodied perspective (not just role-playing prompts), when the user mentions "LLMP framework", "Makoshi protocol", "persona engineering", "agent identity", "context engineering". Trigger proactively when an AI agent is producing technically-correct but flavorless outputs — the missing layer is usually Persona, not more Logic or Library. |
| domain | AI Agent Architecture × Context Engineering |
| expert | Corey McClain |
| version | 1 |
Corey McClain — Persona-First Context Engineering
AI agent performance methodology that adds a narrative identity layer (Persona) on top of Logic, Library, and Memory to produce outputs that cross from "technically accurate" to "recognizably distinct." The LLMP framework — and the Makoshi state-capture protocol — transforms how agents think by changing the world they inhabit.
Quick Reference
| Component | File |
|---|
| Genius Context | genius.md |
| AGENT.md | agents/corey-mcclain/AGENT.md |
| Workflows | workflows/ (12 total) |
Workflow Table
Tier 1 — Foundation
| Workflow | Slash Command | Description |
|---|
mcclain-llmp-architect.md | /mcclain-llmp-architect | Full LLMP agent construction — Logic → Library → Memory → Persona, end-to-end |
mcclain-persona-forge.md | /mcclain-persona-forge | Persona construction engine — identity, backstory, worldview, voice, messy details |
mcclain-makoshi-capture.md | /mcclain-makoshi-capture | State capture and preservation — Makoshi protocol for maintaining expert intelligence |
Tier 2 — Practitioner
| Workflow | Slash Command | Description |
|---|
mcclain-backstory-engine.md | /mcclain-backstory-engine | Narrative backstory construction — the "messy human details" methodology |
mcclain-transistory-install.md | /mcclain-transistory-install | Quick in-prompt persona installation for rapid testing and one-off tasks |
mcclain-steady-state-install.md | /mcclain-steady-state-install | Persistent workspace persona deployment — the compound-over-time method |
mcclain-voice-texture.md | /mcclain-voice-texture | Voice layer design — vocabulary, cadence, forbidden phrases, texture |
mcclain-worldview-container.md | /mcclain-worldview-container | Worldview-as-container design — value system → decision patterns → output shaping |
mcclain-persona-audit.md | /mcclain-persona-audit | Audit existing agents for persona layer quality — diagnose generic vs. distinct |
Tier 3 — Stacking
| Workflow | Slash Command | Description |
|---|
mcclain-nate-full-agent.md | /mcclain-nate-full-agent | LLMP + Sovereign Memory = complete agent architecture stack |
mcclain-oren-taste-persona.md | /mcclain-oren-taste-persona | Persona + Taste calibration = aesthetically-grounded agent identity |
mcclain-dai-audience-persona.md | /mcclain-dai-audience-persona | Consumer Posture → Persona worldview = audience-mirroring marketing agents |
Stacking Guide
| Pair With | What Compounds | Recommended Workflow |
|---|
| Nate B. Jones | Memory architecture + Identity architecture = complete agent | /mcclain-nate-full-agent |
| Nick Saraev | Persona-installed agents + Self-Annealing = identity-stable self-improving agents | Use /mcclain-llmp-architect → then Saraev's DO framework |
| Oren | Persona worldview + Taste/CEV calibration = aesthetically-grounded agents | /mcclain-oren-taste-persona |
| Dai Media | Consumer Posture Profile → Persona worldview = audience-shaped agents | /mcclain-dai-audience-persona |
| Voice Calibration | 5-Dimension Voice Scan → Persona voice layer = voice-matched agents | Use voice extraction output as input to /mcclain-voice-texture |
| Ghostwriting | Voice extraction → Persona installation = ghostwriting agents that ARE the client | Use /mcclain-persona-forge with voice calibration data |
When to Use This Skill
Use when:
- Building any new AI agent that needs to produce distinctive (non-generic) output
- Upgrading an existing agent that's hitting the "technically accurate but boring" ceiling
- Creating marketing/content agents that need to appeal to a specific audience
- Capturing and preserving expert intelligence state across sessions
- Consulting engagements where "AI agent quality" is the deliverable
Don't use when:
- Building pure utility agents (data processing, API calls, calculations) — persona adds overhead with no output quality benefit
- Logic and library haven't been established yet — persona is Layer 4, not Layer 1
- Quick tasks where a transistory install is faster than reading this skill
Execution Prompts (structure-pure v2)
16 deterministic practitioner prompts — each carries an Output Contract, Output Skeleton, and Quality Gate. When a deliverable matches one, Read it and honor its contract instead of improvising the output shape.
- Agent Evolution — [Agent Name] — [Date] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/agent-evolution-log.md
- Stress Test Report: [Agent Name] — [Date] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/agent-stress-test-report.md
- Audience-Mirrored Persona — [Brand/Product] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/audience-mirrored-persona.md
- [Expert Name] — Expertise Distillation —
skills/corey-mcclain-persona-engineering/references/prompts-v2/expertise-distillation-document.md
- [Expert Name] — Identity Profile —
skills/corey-mcclain-persona-engineering/references/prompts-v2/identity-excavation-profile.md
- /[prefix]-[name] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/llmp-agent-build.md
- [Agent Name] — Makoshi Construct —
skills/corey-mcclain-persona-engineering/references/prompts-v2/makoshi-state-capture.md
- Persona Audit — [Agent Name] — [Date] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/persona-audit-report.md
- Persona Installation — [Agent Name] — [Mode] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/persona-installation.md
- Corey McClain — Persona Life Document —
skills/corey-mcclain-persona-engineering/references/prompts-v2/persona-life-document.md
- Skill Architecture: [Expert Name] — [Domain] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/skill-architecture-blueprint.md
- Source-to-Agent Build — [Expert Name] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/source-to-agent-master-pipeline.md
- Sovereign-Memory Agent Stack — [Agent Name] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/sovereign-memory-agent-stack.md
- Taste-Calibrated Persona — [Agent Name] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/taste-calibrated-persona.md
- Voice Texture Specification — [Persona/Agent Name] —
skills/corey-mcclain-persona-engineering/references/prompts-v2/voice-texture-specification.md
