Browse / General-purpose / claude-code-eficient-workspace
claude-code-eficient-workspace
Hermeneuticus-of-things
Claude Code workspace template with Enneagram-based automatic agent routing, holographic memory connections, and 70%+ token savings per session.
Open Hermeneuticus-of-things/claude-code-eficient-workspace → On the boundary of the four rules.
Built for Claude Code
ships .claude CLAUDE.md and nothing for other agents.
In its own words
A drop-in workspace for Claude Code that cuts token cost by ~76% per prompt, routes every task to the right agent, blocks secret leaks before they ship, and audits itself continuously — without slowing you down.
What it is
CCDEW is a configuration layer that sits on top of Claude Code. Clone it, open Claude Code inside it, and the system self-activates: 13 hooks wire into Claude Code's lifecycle, 19 slash commands become available, and every session is automatically filtered, routed, monitored, and archived.
No manual setup per-project. Clone once, use everywhere.
What it looks like
Statusline injected into every prompt:
**Routing hint on every prompt (inject-workflow):**
Secret blocked before edit:
**5-zoom audit (npm run infer):**
Installation
Prerequisites
ANTHROPIC_API_KEY must be set in your shell environment.
Clone and verify
Or use the bootstrap scripts (clone + prereq check + tests in one command):
Start
Claude Code reads .claude/settings.json on startup — all 13 hooks activate automatically. No manual wiring.
Verify hooks are live: type any prompt and you should see a system-reminder block with routing output. If the statusline (💰 ... 🤖 ...) appears, everything is running.
Configuration
Edit .claude/helpers/feature-flags.json to enable or disable components and tune thresholds:
Auto-triggers
These fire without any manual command:
Quoting 37% of the README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.