claude-flow 2.7.13 → 2.7.15
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/bin/claude-flow +1 -1
- package/dist/src/cli/help-formatter.js +5 -3
- package/dist/src/cli/help-formatter.js.map +1 -1
- package/dist/src/cli/simple-cli.js +104 -0
- package/dist/src/cli/simple-cli.js.map +1 -1
- package/dist/src/cli/validation-helper.js.map +1 -1
- package/dist/src/core/version.js +1 -1
- package/dist/src/memory/swarm-memory.js +0 -13
- package/dist/src/memory/swarm-memory.js.map +1 -1
- package/dist/src/utils/metrics-reader.js +41 -29
- package/dist/src/utils/metrics-reader.js.map +1 -1
- package/docs/AGENTIC_FLOW_INTEGRATION_REVIEW.md +593 -0
- package/docs/INTEGRATION_STATUS_FINAL.md +637 -0
- package/docs/LATEST_LIBRARIES_REVIEW.md +937 -0
- package/docs/MEMORY_COMMAND_FIX.md +249 -0
- package/docs/RELEASE_NOTES_v2.7.15.md +332 -0
- package/docs/REMOTE_INSTALL_FIX.md +120 -0
- package/docs/SWARM_INITIALIZATION_GUIDE.md +302 -0
- package/docs/TOOL_VALIDATION_REPORT.md +666 -0
- package/docs/V2.7.14_RELEASE_NOTES.md +102 -0
- package/docs/VALIDATION_REPORT_v2.7.1.md +316 -0
- package/package.json +5 -8
- /package/.claude/{null-settings.json → settings.json} +0 -0
|
@@ -0,0 +1,302 @@
|
|
|
1
|
+
# Swarm Initialization Guide
|
|
2
|
+
**Date:** 2025-10-25
|
|
3
|
+
**Claude-Flow Version:** 2.7.14
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
## ❌ Common Mistake
|
|
8
|
+
|
|
9
|
+
```bash
|
|
10
|
+
# THIS DOES NOT EXIST:
|
|
11
|
+
npx claude-flow@alpha hooks swarm-init --topology adaptive --max-agents 6
|
|
12
|
+
# Error: ❌ Unknown hooks command: swarm-init
|
|
13
|
+
```
|
|
14
|
+
|
|
15
|
+
**Why?** Hooks are for **lifecycle events** (pre-task, post-task, etc.), NOT swarm initialization.
|
|
16
|
+
|
|
17
|
+
---
|
|
18
|
+
|
|
19
|
+
## ✅ Correct Methods
|
|
20
|
+
|
|
21
|
+
### Method 1: MCP Tools (In Claude Code)
|
|
22
|
+
|
|
23
|
+
**For ruv-swarm (WASM-powered):**
|
|
24
|
+
```javascript
|
|
25
|
+
// Call this from within Claude Code:
|
|
26
|
+
mcp__ruv-swarm__swarm_init({
|
|
27
|
+
topology: "mesh", // mesh, hierarchical, ring, star
|
|
28
|
+
maxAgents: 6,
|
|
29
|
+
strategy: "adaptive" // balanced, specialized, adaptive
|
|
30
|
+
})
|
|
31
|
+
```
|
|
32
|
+
|
|
33
|
+
**For claude-flow MCP:**
|
|
34
|
+
```javascript
|
|
35
|
+
// Call this from within Claude Code:
|
|
36
|
+
mcp__claude-flow__swarm_init({
|
|
37
|
+
topology: "hierarchical",
|
|
38
|
+
maxAgents: 8,
|
|
39
|
+
strategy: "balanced"
|
|
40
|
+
})
|
|
41
|
+
```
|
|
42
|
+
|
|
43
|
+
**Example Result:**
|
|
44
|
+
```json
|
|
45
|
+
{
|
|
46
|
+
"id": "swarm-1761410358918",
|
|
47
|
+
"message": "Successfully initialized mesh swarm with 6 max agents",
|
|
48
|
+
"topology": "mesh",
|
|
49
|
+
"strategy": "adaptive",
|
|
50
|
+
"maxAgents": 6,
|
|
51
|
+
"features": {
|
|
52
|
+
"cognitive_diversity": true,
|
|
53
|
+
"neural_networks": true,
|
|
54
|
+
"forecasting": false,
|
|
55
|
+
"simd_support": true
|
|
56
|
+
},
|
|
57
|
+
"performance": {
|
|
58
|
+
"initialization_time_ms": 1.28,
|
|
59
|
+
"memory_usage_mb": 48
|
|
60
|
+
}
|
|
61
|
+
}
|
|
62
|
+
```
|
|
63
|
+
|
|
64
|
+
---
|
|
65
|
+
|
|
66
|
+
### Method 2: Claude-Flow CLI (High-Level)
|
|
67
|
+
|
|
68
|
+
**For task-based swarm deployment:**
|
|
69
|
+
```bash
|
|
70
|
+
# Requires Claude Code CLI
|
|
71
|
+
npx claude-flow@alpha swarm "Build a REST API with authentication" --max-agents 6
|
|
72
|
+
|
|
73
|
+
# With options:
|
|
74
|
+
npx claude-flow@alpha swarm "Analyze codebase" \
|
|
75
|
+
--max-agents 6 \
|
|
76
|
+
--strategy research \
|
|
77
|
+
--mode mesh \
|
|
78
|
+
--read-only
|
|
79
|
+
```
|
|
80
|
+
|
|
81
|
+
**CLI Options:**
|
|
82
|
+
- `--strategy` - research, development, analysis, testing, optimization
|
|
83
|
+
- `--mode` - centralized, distributed, hierarchical, mesh, hybrid
|
|
84
|
+
- `--max-agents` - Maximum number of agents (default: 5)
|
|
85
|
+
- `--parallel` - Enable parallel execution (2.8-4.4x speedup)
|
|
86
|
+
- `--monitor` - Real-time swarm monitoring
|
|
87
|
+
- `--background` - Run in background
|
|
88
|
+
- `--claude` - Open Claude Code CLI
|
|
89
|
+
- `--executor` - Use built-in executor
|
|
90
|
+
- `--read-only` - Analysis mode (no code changes)
|
|
91
|
+
|
|
92
|
+
---
|
|
93
|
+
|
|
94
|
+
### Method 3: Hive Mind System (Recommended for Complex Tasks)
|
|
95
|
+
|
|
96
|
+
**Interactive wizard:**
|
|
97
|
+
```bash
|
|
98
|
+
npx claude-flow@alpha hive-mind wizard
|
|
99
|
+
```
|
|
100
|
+
|
|
101
|
+
**Direct spawn:**
|
|
102
|
+
```bash
|
|
103
|
+
npx claude-flow@alpha hive-mind spawn "Build REST API with auth"
|
|
104
|
+
npx claude-flow@alpha hive-mind spawn "Analyze security" --claude
|
|
105
|
+
```
|
|
106
|
+
|
|
107
|
+
**Check status:**
|
|
108
|
+
```bash
|
|
109
|
+
npx claude-flow@alpha hive-mind status
|
|
110
|
+
npx claude-flow@alpha hive-mind metrics
|
|
111
|
+
```
|
|
112
|
+
|
|
113
|
+
---
|
|
114
|
+
|
|
115
|
+
## 🎯 When to Use Each Method
|
|
116
|
+
|
|
117
|
+
### Use MCP Tools When:
|
|
118
|
+
- ✅ Working inside Claude Code
|
|
119
|
+
- ✅ Need low-level control over swarm topology
|
|
120
|
+
- ✅ Want to integrate with other MCP operations
|
|
121
|
+
- ✅ Building custom coordination workflows
|
|
122
|
+
|
|
123
|
+
### Use CLI Commands When:
|
|
124
|
+
- ✅ Running from terminal/scripts
|
|
125
|
+
- ✅ Need quick swarm deployment
|
|
126
|
+
- ✅ Want high-level task orchestration
|
|
127
|
+
- ✅ Prefer command-line interface
|
|
128
|
+
|
|
129
|
+
### Use Hive Mind When:
|
|
130
|
+
- ✅ Complex multi-objective tasks
|
|
131
|
+
- ✅ Need persistent swarm coordination
|
|
132
|
+
- ✅ Want intelligent agent self-organization
|
|
133
|
+
- ✅ Require SQLite-backed memory
|
|
134
|
+
|
|
135
|
+
---
|
|
136
|
+
|
|
137
|
+
## 🔧 Hooks System (Lifecycle Events)
|
|
138
|
+
|
|
139
|
+
**Hooks are for lifecycle management, NOT swarm creation:**
|
|
140
|
+
|
|
141
|
+
### Available Hook Commands:
|
|
142
|
+
```bash
|
|
143
|
+
# Before task starts
|
|
144
|
+
npx claude-flow@alpha hooks pre-task \
|
|
145
|
+
--description "Build API" \
|
|
146
|
+
--task-id "task-123"
|
|
147
|
+
|
|
148
|
+
# After task completes
|
|
149
|
+
npx claude-flow@alpha hooks post-task \
|
|
150
|
+
--task-id "task-123" \
|
|
151
|
+
--analyze-performance
|
|
152
|
+
|
|
153
|
+
# Before file edit
|
|
154
|
+
npx claude-flow@alpha hooks pre-edit \
|
|
155
|
+
--file "src/api.js" \
|
|
156
|
+
--operation edit
|
|
157
|
+
|
|
158
|
+
# After file edit
|
|
159
|
+
npx claude-flow@alpha hooks post-edit \
|
|
160
|
+
--file "src/api.js" \
|
|
161
|
+
--memory-key "swarm/123/edits/api"
|
|
162
|
+
|
|
163
|
+
# End of session
|
|
164
|
+
npx claude-flow@alpha hooks session-end \
|
|
165
|
+
--export-metrics \
|
|
166
|
+
--generate-summary
|
|
167
|
+
```
|
|
168
|
+
|
|
169
|
+
---
|
|
170
|
+
|
|
171
|
+
## 🌐 Flow-Nexus Cloud Platform
|
|
172
|
+
|
|
173
|
+
**For cloud-based swarm orchestration:**
|
|
174
|
+
|
|
175
|
+
```javascript
|
|
176
|
+
// 1. Register/Login
|
|
177
|
+
mcp__flow-nexus__user_register({ email, password })
|
|
178
|
+
mcp__flow-nexus__user_login({ email, password })
|
|
179
|
+
|
|
180
|
+
// 2. Initialize cloud swarm
|
|
181
|
+
mcp__flow-nexus__swarm_init({
|
|
182
|
+
topology: "hierarchical",
|
|
183
|
+
maxAgents: 8,
|
|
184
|
+
strategy: "balanced"
|
|
185
|
+
})
|
|
186
|
+
|
|
187
|
+
// 3. Spawn cloud agents
|
|
188
|
+
mcp__flow-nexus__agent_spawn({
|
|
189
|
+
type: "researcher",
|
|
190
|
+
name: "researcher-1",
|
|
191
|
+
capabilities: ["web-search", "data-analysis"]
|
|
192
|
+
})
|
|
193
|
+
|
|
194
|
+
// 4. Orchestrate tasks
|
|
195
|
+
mcp__flow-nexus__task_orchestrate({
|
|
196
|
+
task: "Build microservices architecture",
|
|
197
|
+
strategy: "parallel",
|
|
198
|
+
priority: "high"
|
|
199
|
+
})
|
|
200
|
+
```
|
|
201
|
+
|
|
202
|
+
---
|
|
203
|
+
|
|
204
|
+
## 📊 Comparison Matrix
|
|
205
|
+
|
|
206
|
+
| Method | Speed | Control | Complexity | Best For |
|
|
207
|
+
|--------|-------|---------|------------|----------|
|
|
208
|
+
| **MCP Tools** | Fast (1-2ms) | High | Low | Claude Code integration |
|
|
209
|
+
| **CLI Swarm** | Medium | Medium | Medium | Terminal workflows |
|
|
210
|
+
| **Hive Mind** | Slower | High | High | Complex orchestration |
|
|
211
|
+
| **Flow-Nexus** | Variable | Highest | High | Cloud deployment |
|
|
212
|
+
|
|
213
|
+
---
|
|
214
|
+
|
|
215
|
+
## 🚀 Quick Start Examples
|
|
216
|
+
|
|
217
|
+
### Example 1: Simple Research Swarm
|
|
218
|
+
```bash
|
|
219
|
+
npx claude-flow@alpha swarm "Research GraphQL best practices" \
|
|
220
|
+
--strategy research \
|
|
221
|
+
--max-agents 3 \
|
|
222
|
+
--read-only
|
|
223
|
+
```
|
|
224
|
+
|
|
225
|
+
### Example 2: Development Swarm
|
|
226
|
+
```bash
|
|
227
|
+
npx claude-flow@alpha swarm "Build authentication service" \
|
|
228
|
+
--strategy development \
|
|
229
|
+
--max-agents 5 \
|
|
230
|
+
--parallel \
|
|
231
|
+
--monitor
|
|
232
|
+
```
|
|
233
|
+
|
|
234
|
+
### Example 3: MCP-Based Swarm
|
|
235
|
+
```javascript
|
|
236
|
+
// In Claude Code:
|
|
237
|
+
mcp__ruv-swarm__swarm_init({
|
|
238
|
+
topology: "mesh",
|
|
239
|
+
maxAgents: 4,
|
|
240
|
+
strategy: "adaptive"
|
|
241
|
+
})
|
|
242
|
+
|
|
243
|
+
mcp__ruv-swarm__agent_spawn({
|
|
244
|
+
type: "coder",
|
|
245
|
+
name: "backend-dev"
|
|
246
|
+
})
|
|
247
|
+
|
|
248
|
+
mcp__ruv-swarm__task_orchestrate({
|
|
249
|
+
task: "Implement user authentication",
|
|
250
|
+
strategy: "sequential",
|
|
251
|
+
priority: "high"
|
|
252
|
+
})
|
|
253
|
+
```
|
|
254
|
+
|
|
255
|
+
### Example 4: Hive Mind with Claude Code
|
|
256
|
+
```bash
|
|
257
|
+
npx claude-flow@alpha hive-mind spawn \
|
|
258
|
+
"Build e-commerce platform" \
|
|
259
|
+
--claude
|
|
260
|
+
```
|
|
261
|
+
|
|
262
|
+
---
|
|
263
|
+
|
|
264
|
+
## 🐞 Troubleshooting
|
|
265
|
+
|
|
266
|
+
### Issue: "Unknown hooks command: swarm-init"
|
|
267
|
+
**Solution:** Use `npx claude-flow@alpha swarm` or MCP tools instead
|
|
268
|
+
|
|
269
|
+
### Issue: "Compiled swarm module not found"
|
|
270
|
+
**Solution:** Either:
|
|
271
|
+
1. Use MCP tools (recommended)
|
|
272
|
+
2. Install Claude Code CLI
|
|
273
|
+
3. Use `--executor` flag for built-in execution
|
|
274
|
+
|
|
275
|
+
### Issue: Swarm initialization fails
|
|
276
|
+
**Check:**
|
|
277
|
+
1. MCP servers running: `claude mcp list`
|
|
278
|
+
2. Dependencies installed: `npm install`
|
|
279
|
+
3. Memory available: Check `.swarm/memory.db`
|
|
280
|
+
|
|
281
|
+
---
|
|
282
|
+
|
|
283
|
+
## 📚 Additional Resources
|
|
284
|
+
|
|
285
|
+
**Documentation:**
|
|
286
|
+
- Main README: `/README.md`
|
|
287
|
+
- Tool Validation: `/docs/TOOL_VALIDATION_REPORT.md`
|
|
288
|
+
- Integration Review: `/docs/AGENTIC_FLOW_INTEGRATION_REVIEW.md`
|
|
289
|
+
|
|
290
|
+
**Skills:**
|
|
291
|
+
- Swarm Orchestration: `.claude/skills/swarm-orchestration/SKILL.md`
|
|
292
|
+
- Hive Mind Advanced: `.claude/skills/hive-mind-advanced/SKILL.md`
|
|
293
|
+
|
|
294
|
+
**Official Guides:**
|
|
295
|
+
- GitHub: https://github.com/ruvnet/claude-flow
|
|
296
|
+
- Hive Mind: https://github.com/ruvnet/claude-flow/tree/main/docs/hive-mind
|
|
297
|
+
|
|
298
|
+
---
|
|
299
|
+
|
|
300
|
+
**Created:** 2025-10-25
|
|
301
|
+
**Updated:** 2025-10-25
|
|
302
|
+
**Author:** Claude Code (Claude Sonnet 4.5)
|