Features
Context Meter
See how much of your context window you've used at a glance:
█░░░░░░░░░ 8% ← Fresh
████░░░░░░ 40% ← Getting going
███████░░░ 65% ← Consider wrapping up
██████████ 95% ← Autocompact incomingColor-coded: green → yellow → orange → blinking red.
AI Asset Discovery
As you type, the status bar shows the most relevant installed AI assets:
You type: "help me do SEO"
Status: ... │ 💡 /marketing:seo-audit /backlink-builder
You type: "debug this API error"
Status: ... │ 💡 /gsd-debug Agent("gsd-debugger")
You type: "做个数据库查询"
Status: ... │ 💡 /data:write-query /data:sql-queriesScans all asset types:
- Skills (
~/.claude/skills/) - Agents (
~/.claude/agents/) - Commands (
~/.claude/commands/) - Plugins (Claude Code built-in)
- MCP Servers (
.mcp.json)
Smart matching — not just keyword search:
- 35+ intent categories with Chinese + English aliases
- TF-IDF weighting (rare keywords matter more)
- Specificity bonus (specialist tools rank above generalists)
- Diversity filter (max 2 per namespace)
How It Works
You type a message
│
▼
UserPromptSubmit hook
→ Tokenize (CJK bigrams + Latin words)
→ Detect intents (35+ categories)
→ Score assets (TF-IDF + specificity)
→ Write top 3 to temp bridge file
│
▼
Statusline reads bridge file
→ Display: model │ dir │ context% │ 💡 assetsZero token consumption. Everything runs as local Node.js processes.
FAQ
Q: What is CC Status Board? A: A smart status bar extension for Claude Code that shows context usage, model info, and intelligently suggests relevant installed AI assets based on your conversation. Open source, MIT licensed.
Q: Does it consume tokens? A: No. Zero token consumption. All matching runs 100% locally as Node.js processes.
Q: How do I install CC Status Board?
A: Clone the repo and run node install.js, or use npx cc-status-board. Restart Claude Code after installation.
Q: What assets does it scan? A: Skills, agents, commands, Claude Code plugins, and MCP servers — typically 200-400 assets for active users.