feat: GoZen v3.0.0 - Context Compression & Middleware Pipeline

.github/workflows/ci.yml

@@ -12,6 +12,20 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - uses: actions/setup-node@v4
+        with:
+          node-version: 22
+
+      - name: Build Web UI
+        working-directory: web
+        run: |
+          npm ci
+          npm run build
+
+      - name: Test Web UI
+        working-directory: web
+        run: npm run test:coverage
+
       - name: Read Go version
         id: goversion
         run: echo "version=$(grep '^go ' go.mod | awk '{print $2}')" >> "$GITHUB_OUTPUT"
@@ -29,6 +43,9 @@ jobs:
       - name: Test
         run: go test -race -count=1 ./...
 
+      - name: Integration Tests
+        run: go test -tags=integration -race -count=1 -timeout 180s ./test/integration/...
+
       - name: Check coverage thresholds
         run: |
           check_pkg() {
@@ -53,7 +70,8 @@ jobs:
           check_pkg "internal/proxy" 80 || failed=1
           check_pkg "internal/proxy/transform" 80 || failed=1
           check_pkg "internal/web" 80 || failed=1
-          check_pkg "internal/daemon" 40 || failed=1
+          check_pkg "internal/bot" 80 || failed=1
+          check_pkg "internal/daemon" 50 || failed=1
           check_pkg "internal/update" 50 || failed=1
           check_pkg "internal/sync" 50 || failed=1
 

.github/workflows/release.yml

@@ -50,6 +50,16 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - uses: actions/setup-node@v4
+        with:
+          node-version: 22
+
+      - name: Build Web UI
+        working-directory: web
+        run: |
+          npm ci
+          npm run build
+
       - uses: actions/setup-go@v5
         with:
           go-version-file: go.mod

.gitignore

@@ -5,3 +5,8 @@
 /dist/
 /temp-docs/
 /.dev/
+web/node_modules/
+web/dist/
+internal/web/dist/
+bin/
+web/coverage/

CLAUDE.md

@@ -33,6 +33,25 @@ go build ./...
 go test ./...
 ```
 
+## Dev Environment
+
+Dev daemon runs on separate ports to avoid interfering with production:
+
+- **Dev Web UI**: `http://localhost:29840`
+- **Dev Proxy**: `http://localhost:29841`
+- **Dev Config**: `~/.zen-dev/zen.json`
+- **Production ports (19840/19841)**: NEVER touch during development
+
+```sh
+./scripts/dev.sh              # Start dev daemon (builds + starts)
+./scripts/dev.sh stop         # Stop dev daemon
+./scripts/dev.sh restart      # Rebuild and restart dev daemon
+./scripts/dev.sh status       # Check dev daemon status
+./scripts/dev.sh web          # Start frontend dev server (Vite)
+```
+
+After modifying Go code, always run `./scripts/dev.sh restart` to rebuild and restart the dev daemon.
+
 ## Release Process
 
 Push a git tag to trigger GitHub Actions release workflow:
@@ -56,6 +75,24 @@ Do NOT use `gh release create` — the CI pipeline handles release creation auto
 - **No example files**: Do NOT create example config files (*.json, *.yaml, etc.) in the repository root. Examples belong in README.md or `docs/`.
 - **Minimal test files**: Only add tests for new public APIs or complex logic. Do not create excessive test files for simple functions. Prefer table-driven tests in existing *_test.go files.
 - **No unnecessary files**: Before committing, review `git status` and remove any generated, temporary, or example files that should not be in the repository.
+- **TDD for new features**: Use Test-Driven Development (TDD) for new feature development to ensure code quality. Write tests first, then implement the feature to make tests pass.
+
+## Release Checklist
+
+Before tagging a release, complete the following checklist:
+
+1. **Bug check**: Review all code for unresolved bugs. Run `go test ./...` and ensure all tests pass.
+2. **Version number**: Verify `Version` in `cmd/root.go` matches the release tag.
+3. **Website documentation**:
+   - Ensure website contains documentation for the release version
+   - Verify documentation is accurate and complete
+   - Confirm all new features are documented
+   - Remove or update documentation for changed/removed features
+4. **README files**: Update all README files to reflect the latest version:
+   - `README.md` (English)
+   - `docs/README.zh-CN.md` (简体中文)
+   - `docs/README.zh-TW.md` (繁體中文)
+   - `docs/README.es.md` (Español)
 
 ## Config Migration Rules
 
@@ -108,3 +145,4 @@ Background (Light): `#f8fafc` → `#ffffff` → `#f1f5f9` → `#e2e8f0`
 - v1.5.2: Allow reinstalling same version in upgrade command
 - v1.5.3: Per-binary PID files to avoid multi-binary conflicts
 - v2.0.0: Rename to GoZen (opencc → zen), config migration from ~/.opencc/ to ~/.zen/, non-blocking version update check
+- v3.0.0: Usage tracking & budget control, provider health monitoring, smart load balancing, webhooks, context compression, middleware pipeline, agent infrastructure

README.md

@@ -26,6 +26,18 @@ Multi-CLI environment switcher for Claude Code, Codex, and OpenCode with API pro
 - **Self-Update** — One-command upgrade via `zen upgrade` with semver version matching (supports prerelease versions)
 - **Shell Completion** — Supports zsh / bash / fish
 
+### v3.0 New Features
+
+- **Usage Tracking** — Track token usage and costs per provider, model, and project
+- **Budget Control** — Set daily/weekly/monthly spending limits with warn/downgrade/block actions
+- **Provider Health Monitoring** — Real-time health checks with latency and error rate tracking
+- **Smart Load Balancing** — Multiple strategies: failover, round-robin, least-latency, least-cost
+- **Webhooks** — Notifications for budget alerts, provider status changes, and daily summaries (Slack/Discord/Generic)
+- **Context Compression** — Automatic context compression when token count exceeds threshold
+- **Middleware Pipeline** — Pluggable middleware for request/response transformation
+- **Agent Infrastructure** — Built-in support for agent-based workflows with session management
+- **Bot Gateway** — Monitor and control Claude Code sessions remotely via Telegram, Discord, Slack, Lark, or Facebook Messenger
+
 ## Installation
 
 ```sh
@@ -89,7 +101,7 @@ zen --cli codex
 
 ## Daemon Architecture
 
-In v2.1, GoZen uses a unified daemon process (`zend`) that hosts both the HTTP proxy and the Web UI:
+In v3.0, GoZen uses a unified daemon process (`zend`) that hosts both the HTTP proxy and the Web UI:
 
 - **Proxy server** runs on port `19841` (configurable via `proxy_port`)
 - **Web UI** runs on port `19840` (configurable via `web_port`)
@@ -330,6 +342,176 @@ Configuration example:
 }
 ```
 
+## Usage Tracking & Budget Control
+
+Track API usage and set spending limits:
+
+```json
+{
+  "pricing": {
+    "claude-sonnet-4-20250514": {"input_per_million": 3.0, "output_per_million": 15.0},
+    "claude-opus-4-20250514": {"input_per_million": 15.0, "output_per_million": 75.0}
+  },
+  "budgets": {
+    "daily": {"amount": 10.0, "action": "warn"},
+    "monthly": {"amount": 100.0, "action": "block"},
+    "per_project": true
+  }
+}
+```
+
+Budget actions: `warn` (log warning), `downgrade` (switch to cheaper model), `block` (reject requests).
+
+## Provider Health Monitoring
+
+Automatic health checks with metrics tracking:
+
+```json
+{
+  "health_check": {
+    "enabled": true,
+    "interval_secs": 60,
+    "timeout_secs": 10
+  }
+}
+```
+
+View provider health via Web UI or API: `GET /api/v1/health/providers`
+
+## Smart Load Balancing
+
+Configure load balancing strategy per profile:
+
+```json
+{
+  "profiles": {
+    "balanced": {
+      "providers": ["provider-a", "provider-b", "provider-c"],
+      "strategy": "least-latency"
+    }
+  }
+}
+```
+
+Strategies:
+- `failover` — Try providers in order (default)
+- `round-robin` — Distribute requests evenly
+- `least-latency` — Route to fastest provider
+- `least-cost` — Route to cheapest provider for the model
+
+## Webhooks
+
+Get notified about important events:
+
+```json
+{
+  "webhooks": [
+    {
+      "name": "slack-alerts",
+      "url": "https://hooks.slack.com/services/xxx",
+      "events": ["budget_warning", "budget_exceeded", "provider_down", "provider_up"],
+      "enabled": true
+    }
+  ]
+}
+```
+
+Events: `budget_warning`, `budget_exceeded`, `provider_down`, `provider_up`, `failover`, `daily_summary`
+
+## Context Compression
+
+Automatically compress context when it exceeds a threshold:
+
+```json
+{
+  "compression": {
+    "enabled": true,
+    "threshold_tokens": 100000,
+    "target_ratio": 0.5
+  }
+}
+```
+
+## Middleware Pipeline
+
+Transform requests and responses with pluggable middleware:
+
+```json
+{
+  "middleware": {
+    "enabled": true,
+    "middlewares": [
+      {"name": "context-injection", "enabled": true, "config": {"inject_cursorrules": true}},
+      {"name": "rate-limiter", "enabled": true, "config": {"requests_per_minute": 60}}
+    ]
+  }
+}
+```
+
+Built-in middleware: `context-injection`, `request-logger`, `rate-limiter`, `compression`
+
+## Bot Gateway
+
+Monitor and control your Claude Code sessions remotely via chat platforms. The bot connects to running `zen` processes via IPC and lets you:
+
+- View connected processes and their status
+- Send tasks to specific processes
+- Receive notifications for approvals, errors, and completions
+- Control tasks (pause/resume/cancel)
+
+### Supported Platforms
+
+| Platform | Setup Required |
+|----------|----------------|
+| Telegram | BotFather token |
+| Discord | Bot application token |
+| Slack | Bot + App tokens (Socket Mode) |
+| Lark/Feishu | App ID + Secret |
+| Facebook Messenger | Page token + Verify token |
+
+### Bot Commands
+
+| Command | Description |
+|---------|-------------|
+| `list` | List all connected processes |
+| `status [name]` | Show process status |
+| `bind <name>` | Bind to a process for subsequent commands |
+| `pause/resume/cancel [name]` | Control tasks |
+| `<name> <task>` | Send a task to a process |
+| `help` | Show available commands |
+
+The bot also understands natural language queries like "帮我看看 gozen 的状态" (show me gozen's status).
+
+### Configuration
+
+```json
+{
+  "bot": {
+    "enabled": true,
+    "socket_path": "/tmp/zen-bot.sock",
+    "platforms": {
+      "telegram": {
+        "enabled": true,
+        "token": "123456:ABC-DEF...",
+        "allowed_users": ["your_username"]
+      }
+    },
+    "interaction": {
+      "require_mention": true,
+      "mention_keywords": ["@zen", "/zen"],
+      "direct_message_mode": "always",
+      "channel_mode": "mention"
+    },
+    "aliases": {
+      "api": "/path/to/api-project",
+      "web": "/path/to/web-project"
+    }
+  }
+}
+```
+
+See the [Bot Gateway documentation](https://gozen.dev/docs/bot) for detailed platform setup guides.
+
 ## Config Files
 
 | File | Description |
@@ -343,7 +525,7 @@ Configuration example:
 
 ```json
 {
-  "version": 7,
+  "version": 8,
   "default_profile": "default",
   "default_client": "claude",
   "proxy_port": 19841,
@@ -379,11 +561,11 @@ Configuration example:
 zen upgrade
 
 # Specific version
-zen upgrade 2.1
-zen upgrade 2.1.0
+zen upgrade 3.0
+zen upgrade 3.0.0
 
 # Prerelease version
-zen upgrade 2.1.0-alpha.1
+zen upgrade 3.0.0-alpha.1
 ```
 
 ## Migrating from Older Versions
@@ -405,8 +587,8 @@ go test ./...
 Release: Push a tag and GitHub Actions will build automatically.
 
 ```sh
-git tag v2.1.0
-git push origin v2.1.0
+git tag v3.0.0
+git push origin v3.0.0
 ```
 
 ## License