━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━


VIBE CODER SCORE


Level 64 · Commander

"Managing a parallel workforce"


Deep tool ecosystem. Custom MCP servers, comprehensive coverage — you're a producer, not just a consumer. Social is your biggest growth area. 7 MCP servers connected — that's serious tool integration. Pioneer badge earned — you're building what the taxonomy hasn't seen yet.


Intelligence ████████░░ 80

Tooling ██████████ 100

Continuity ███████░░░ 65

Autonomy ███████░░░ 65

Ship ███████░░░ 70

Security █████░░░░░ 45

Ops ██░░░░░░░░ 20

Social ██░░░░░░░░ 15


WHAT THE SCAN FOUND

────────────────────────────────────────────────────


Intelligence

Anthropic (Claude) [adv]

OpenAI (GPT) [inter]

Google Gemini [inter]

Model routing aliases [adv]

Tooling

Claude Code CLI [adv]

Cursor IDE [inter]

MCP: Filesystem [inter]

MCP: GitHub [inter]

MCP: Obsidian [inter]

MCP: Supabase [adv]

MCP: Google Drive [inter]

MCP: Brave Search [inter]

MCP: Voice Mode [adv] *

Continuity

Detailed CLAUDE.md [adv]

Claude Code memories [inter]

Split rules directory [adv]

Autonomy

Claude Code subagents [adv]

Claude Code hooks [adv]

tmux configuration [inter]

Ship

Vercel deployment [inter]

Netlify deployment [inter]

GitHub Actions CI [adv]

Test framework (Vitest) [inter]

Security

.gitignore covers .env [basic]

Environment variable secrets [inter]

Agent permission scoping [adv]

Ops

Task tracking in CLAUDE.md [inter]

Build/dev/lint scripts [basic]

Social

Public GitHub repositories [inter]


29 signals · 8 categories

* = novel detection (not in standard registry)


★ Pioneer Badge

1 innovation(s) detected

→ MCP: Voice Mode


GROWTH AREAS

────────────────────────────────────────────────────


🌐 Social (15/100)

Solo setup. No community integration or multi-channel presence.

📊 Ops (20/100)

Minimal operational infrastructure. Documentation and project management are informal.

🛡️ Security (45/100)

Good hygiene. Gitignored secrets, environment variables, some permission scoping.


DETAILED ANALYSIS

────────────────────────────────────────────────────


You're running three model providers (Anthropic, OpenAI, Gemini) with routing aliases — that means you can dispatch to the right model per task instead of defaulting to one. The 7 MCP servers give Claude Code direct access to GitHub, Supabase, Google Drive, Obsidian, and Brave Search without leaving the terminal. Pair that with custom subagents and hooks, and you've basically built an autonomous pipeline where Claude can research, write, deploy, and document without you context-switching. The split rules directory and 120-line CLAUDE.md show you're investing in persistence — your agent remembers project conventions across sessions, which is where most setups fall short. Worth looking at: adding a pre-commit hook via Claude hooks to run lint/typecheck automatically, wiring up an MCP server for Vercel or Netlify so deploys happen in-conversation, and setting up structured logging in your CI so Claude can self-diagnose failed builds.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━


Get your own score: npx vibecheck-score