Digital Clone Skill
AliceLJY/digital-clone-skill
Crie clones digitais que falam como você a partir do seu corpus de dados
Ver no GitHubSkill para Claude Code que converte histórico de conversas e escritos em um clone digital funcional. Guia o usuário através de 6 estágios — desde coleta e limpeza de corpus até extração de personalidade, geração de system prompt e verificação — com dois modos: Self Mode (clonar a si mesmo) e Mentor Mode (clonar figuras públicas).
Instalação
npx convez-ai AliceLJY/digital-clone-skill1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Criar um assistente digital personalizado baseado em suas conversas e escritos
- Clonar a personalidade de mentores ou figuras públicas para orientação
- Gerar system prompts precisos a partir de análise de voz e comportamento
Veredito

Ferramenta ambiciosa e bem-estruturada para síntese de personalidade em IA. Excelente para quem quer criar assistentes especializados, mas requer disciplina na coleta e limpeza de corpus — a qualidade do clone depende diretamente da qualidade dos dados de entrada.
README
A tradução para o português deste README ainda está na fila.
Digital Clone
Corpus-Driven Digital Clone Skill for Claude Code
Collect your AI conversations, extract your personality, deploy a clone that talks like you.
A Claude Code Skill that turns conversation history and writings into a digital clone — guiding you through corpus collection, cleaning, personality extraction, system prompt generation, and verification. Optional Bun-based CLI/MCP tools are included for mechanical data preprocessing.
English | 简体中文
What It Does
The Skill walks you through a 6-stage pipeline, entirely conversational — no runtime dependencies required:
| Stage | Name | What Happens |
|---|---|---|
| 1 | Target Profiling | Identify the clone target and map data sources |
| 2 | Data Hunting | Collect raw corpus (transcripts, articles, research) |
| 3 | Data Refining | Clean, dedup, PII sanitization, quality assessment |
| 4 | Soul Forging | Extract personality, generate System Prompt |
| 5 | Verification | Trap-question testing with pass criteria (target: ≥80%) |
| 6 | Deployment | Platform-specific deploy guide (NotebookLM / bot / generic LLM) |
Two Modes
- Self Mode — clone yourself from local AI conversations and writings
- Mentor Mode — clone a public figure via 6-angle parallel research (primary voice, live reactions, external views, decisions, social fragments, timeline)
Quick Start
Install the Skill (this is the whole installation):
mkdir -p ~/.claude/skills/digital-clone
curl -o ~/.claude/skills/digital-clone/SKILL.md \
https://raw.githubusercontent.com/AliceLJY/digital-clone-skill/main/SKILL.md
Then in Claude Code:
帮我克隆自己 / "Clone myself from my articles and CC transcripts" 帮我克隆纳瓦尔做数字导师 / "Clone Naval as my digital mentor"
The Skill handles everything conversationally, stage by stage, with your approval at each step. All outputs go to ./clone-workspace/ in your current directory.
Optional: CLI Preprocessing Tools
Requires Bun. The CLI does not run on Node.js (it uses Bun's TypeScript module resolution). If you don't use Bun, skip this section entirely — the Skill covers the full pipeline on its own.
For large corpora (thousands of transcript files), the CLI does the mechanical work faster than in-conversation processing:
git clone https://github.com/AliceLJY/digital-clone-skill.git
cd digital-clone-skill
bun install
bun run src/cli.ts init --target "Your Name" --mode self
bun run src/cli.ts ingest --source all
bun run src/cli.ts refine
bun run src/cli.ts quality
Important: the workspace path is relative to where you run the commands. If you preprocess with the CLI, start your Claude Code session in the same directory so the Skill finds ./clone-workspace/. The refined corpus separates *-user.md (your voice — used for personality extraction) from *-assistant.md (AI replies — reference only, excluded from Soul Forging).
| Command | Description |
|---|---|
bun run src/cli.ts init | Initialize workspace and config |
bun run src/cli.ts ingest --source <src> | Scan corpus (cc, codex, gemini, memory, articles, all) |
bun run src/cli.ts import <path> | Import external files (Mentor Mode) |
bun run src/cli.ts refine | Clean, dedup, sanitize |
bun run src/cli.ts quality | Generate quality report |
bun run src/cli.ts stats | Show corpus statistics |
bun run src/cli.ts verify-template | Generate test case template |
bun run src/cli.ts deploy-guide --platform <p> | Generate deployment guide |
bun run src/cli.ts refresh | Re-scan sources and merge new content into the refined corpus |
Set CLONE_WORKSPACE to pin the workspace to a fixed path shared between CLI and Skill sessions.
refreshcan optionally pull recent memories from a RecallNest install (the author's memory system; setRECALLNEST_CLIor place it at~/recallnest/lm). Without it, use--skip-recallnest.
| Tool | Description |
|---|---|
clone_ingest | Scan and collect corpus |
clone_refine | Clean and deduplicate |
clone_quality | Assess corpus quality |
clone_stats | Show statistics |
clone_read_corpus | Read refined corpus slices (defaults to user-side text) |
MCP Setup (Claude Code):
{
"mcpServers": {
"digital-clone": {
"command": "bun",
"args": ["run", "/path/to/digital-clone-skill/src/mcp-server.ts"],
"cwd": "/path/to/digital-clone-skill"
}
}
}
| File | Role |
|---|---|
SKILL.md | Claude Code Skill — the full 6-stage pipeline (the product) |
src/cli.ts | Optional CLI entry (Bun) |
src/mcp-server.ts | Optional MCP tools (Bun) |
src/parsers.ts | Multi-source transcript parsing |
src/ingest.ts | Corpus collection pipeline |
src/refine.ts | Dedup + PII sanitize + normalize |
src/quality.ts | Quality assessment + report |
src/templates.ts | Verify + deploy template generation |
src/config.ts | Configuration management |
Credit
| Source | Contribution |
|---|---|
| Claude Code | Foundation, CLI, MCP server, parsers |
| RecallNest | Parser architecture for CC/Codex/Gemini transcripts |
| @MinLiBuilds | Naval clone tutorial — original inspiration |
| alchaincyf/nuwa-skill | 6-angle research + three-pass verification |
| LvPengfei1/PersonaVault | Evidence grading + capability boundaries |
Author
Built by 小试AI (@AliceLJY) for the WeChat public account 我的AI小木屋.
Ecosystem
Part of the 小试AI open-source AI workflow:
| Project | Description |
|---|---|
| recallnest | MCP memory workbench (LanceDB + Jina v5) |
| content-publisher | Image generation + layout + WeChat publishing |
| openclaw-tunnel | Docker ↔ host CLI bridge (/cc /codex /gemini) |
| telegram-ai-bridge | Telegram bots for Claude, Codex, and Gemini |
| claude-code-studio | Multi-session collaboration platform for Claude Code |
| cc-empire | Complete Claude Code workflow scaffold (rules + hooks + agents) |
| etwin-bot | E-Twin Telegram bot — this skill's 1:1 instantiation as a runnable bot |
| trio-handoff | Bidirectional handoff bundles for AI coding agents |
License
MIT