me-md
Build a ME.MD profile that captures who you are and how you decide — the personal counterpart to CLAUDE.md, AGENT.md, and DESIGN.md.
me-md is a Claude Code plugin that interviews you, reads your CliftonStrengths report (if you have one), analyzes your real prompt history across Claude Code / Codex / OpenCode, and synthesizes everything into a single markdown file: ME.MD.
Drop the resulting file in ~/ME.MD (or any project root) and import it from your global CLAUDE.md:
## Personal profile
@~/ME.MD
Now every AI session in every project starts with a clear picture of you — your decision style, your quality bar, your stress signals, your tooling stack — instead of you having to re-explain yourself in every new chat.
What's in a ME.MD?
A 12-section structured markdown file with YAML frontmatter:
- Top-line identity — name, mode, default stance, top strengths
- Operating principles — your real, observable rules of engagement
- How you decide — direct imperative vs option-menu, escalation ladder when you disagree
- Style & voice — sentence structure, signature typos, stress tells
- Tooling stack — daily drivers, defaults for new projects, anti-stack
- Work cadence — bursty vs steady, hours, sprint patterns
- Domains you return to — gravitational centers in your work
- Values & quality bar — non-negotiables, in priority order
- How strengths map to behavior — Clifton-style interpretation
- Rules for AI agents reading this file — load-bearing operational guidance
- What this file is not — scope boundaries
- Changelog — versioned, with provenance
Each ME.MD is built with direct quotes from your real prompts as evidence. No invented traits. No flattery.
How it works
/me-md build
The orchestrator walks you through:
- Consent — explains exactly what it's about to read on your machine.
- Assessment detection — asks if you have a CliftonStrengths PDF (or similar). If yes, parses it. If no, runs a structured 20–30 minute interview.
- Source detection — auto-detects which AI tools you have installed (Claude Code, Codex, OpenCode) and asks which to include.
- Parallel analysis — dispatches one subagent per source. Each agent extracts: prompt volume, voice, opening moves, decision signals, top projects, frustrations, signature quirks, time-of-day pattern, and the top quotes that capture how you write.
- Synthesis — combines assessment + log analysis + (optional) interview answers into a draft ME.MD.
- Review — shows you the draft, lets you edit before saving.
- Install — optionally symlinks
~/ME.MD and adds the import line to your global CLAUDE.md.
Run time: 5–10 minutes if you have a CliftonStrengths report; 25–40 minutes with the full interview.
Two modes
Mode A — You have a CliftonStrengths report
Pass the PDF path:
/me-md build path/to/your-clifton-report.pdf
The plugin uses the parse-clifton skill to extract your top themes and uses them as personality scaffolding. Then it triangulates against your actual prompt history.
Mode B — You don't have an assessment
Just run:
/me-md build
The plugin runs the interview-strengths skill, which dispatches a personality-interviewer subagent that conducts a one-question-at-a-time conversational interview covering 8 dimensions:
- Daily energy & cadence
- Decision style
- Learning vs shipping
- Solo vs collaborative
- Detail vs big-picture
- Conflict & feedback
- Tooling promiscuity
- Quality bar & dealbreakers
The interview is honest, behavior-anchored, and explicitly not a validated psychometric instrument — it's a structured way to surface signals that the synthesizer can combine with observed behavior from your logs.
Privacy
- Everything stays on your machine. No prompts, no answers, no logs leave your computer.
- The plugin shows you what it's about to read before reading it.
- You can opt out of any source. ("Skip Codex.")
- No telemetry, no usage reporting.
- The output
ME.MD is in .gitignore by default. You decide if and where to commit it.
See PRIVACY.md (shipped with the plugin) for details.
Components
| Type | Name | Purpose |
|---|
| Skill | build-me-md | Orchestrator and entry point |
| Skill | analyze-prompts | Detects installed AI tools, dispatches log analysis |
| Skill | parse-clifton | Parses CliftonStrengths PDF reports |
| Skill | interview-strengths | Runs the structured interview |
| Skill | synthesize-profile | Combines all data into the final ME.MD |
| Agent | prompt-analyzer | Heavy log analysis (one per source, parallel) |
| Agent | personality-interviewer | Conducts the conversational interview |
Installation
Via the imehr marketplace
Gerstep
dadwadw233
BayramAnnakov
daypunk
mohitagw15856
study8677