agentic-flow 2.0.0-alpha → 2.0.1-alpha
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/README.md +320 -23
- package/agentic-flow/.claude/agents/base-template-generator.md +229 -3
- package/agentic-flow/.claude/agents/core/coder.md +212 -7
- package/agentic-flow/.claude/agents/core/planner.md +228 -7
- package/agentic-flow/.claude/agents/core/researcher.md +205 -10
- package/agentic-flow/.claude/agents/core/reviewer.md +216 -5
- package/agentic-flow/.claude/agents/core/tester.md +213 -3
- package/agentic-flow/.claude/agents/data/ml/data-ml-model.md +256 -5
- package/agentic-flow/.claude/agents/development/backend/dev-backend-api.md +209 -6
- package/agentic-flow/.claude/agents/documentation/api-docs/docs-api-openapi.md +185 -5
- package/agentic-flow/.claude/agents/github/code-review-swarm.md +307 -468
- package/agentic-flow/.claude/agents/github/issue-tracker.md +270 -13
- package/agentic-flow/.claude/agents/github/pr-manager.md +259 -12
- package/agentic-flow/.claude/agents/github/release-manager.md +253 -15
- package/agentic-flow/.claude/agents/github/workflow-automation.md +277 -9
- package/agentic-flow/.claude/agents/sona/sona-learning-optimizer.md +496 -0
- package/agentic-flow/.claude/agents/sparc/architecture.md +231 -4
- package/agentic-flow/.claude/agents/sparc/pseudocode.md +206 -4
- package/agentic-flow/.claude/agents/sparc/refinement.md +283 -6
- package/agentic-flow/.claude/agents/sparc/specification.md +205 -3
- package/agentic-flow/.claude/agents/swarm/adaptive-coordinator.md +731 -0
- package/agentic-flow/.claude/agents/swarm/hierarchical-coordinator.md +455 -1
- package/agentic-flow/.claude/agents/swarm/mesh-coordinator.md +571 -0
- package/agentic-flow/.claude/agents/templates/sparc-coordinator.md +336 -5
- package/agentic-flow/dist/cli/commands/sona-train.d.ts.map +1 -0
- package/agentic-flow/dist/cli/commands/sona-train.js +295 -0
- package/agentic-flow/dist/cli/commands/sona-train.js.map +1 -0
- package/agentic-flow/dist/cli/commands/sona.d.ts.map +1 -0
- package/agentic-flow/dist/cli/commands/sona.js +290 -0
- package/agentic-flow/dist/cli/commands/sona.js.map +1 -0
- package/agentic-flow/dist/core/agentdb-fast.d.ts.map +1 -0
- package/agentic-flow/dist/core/agentdb-fast.js +299 -0
- package/agentic-flow/dist/core/agentdb-fast.js.map +1 -0
- package/agentic-flow/dist/core/attention-fallbacks.d.ts.map +1 -0
- package/agentic-flow/dist/core/attention-fallbacks.js +321 -0
- package/agentic-flow/dist/core/attention-fallbacks.js.map +1 -0
- package/agentic-flow/dist/core/embedding-service.d.ts.map +1 -0
- package/agentic-flow/dist/core/embedding-service.js +370 -0
- package/agentic-flow/dist/core/embedding-service.js.map +1 -0
- package/agentic-flow/dist/core/gnn-wrapper.d.ts.map +1 -0
- package/agentic-flow/dist/core/gnn-wrapper.js +236 -0
- package/agentic-flow/dist/core/gnn-wrapper.js.map +1 -0
- package/agentic-flow/dist/core/index.d.ts.map +1 -1
- package/agentic-flow/dist/core/index.js +80 -3
- package/agentic-flow/dist/core/index.js.map +1 -1
- package/agentic-flow/dist/mcp/claudeFlowSdkServer.d.ts.map +1 -1
- package/agentic-flow/dist/mcp/claudeFlowSdkServer.js +109 -0
- package/agentic-flow/dist/mcp/claudeFlowSdkServer.js.map +1 -1
- package/agentic-flow/dist/mcp/tools/agent-booster-tools.d.ts.map +1 -0
- package/agentic-flow/dist/mcp/tools/agent-booster-tools.js +262 -0
- package/agentic-flow/dist/mcp/tools/agent-booster-tools.js.map +1 -0
- package/agentic-flow/dist/mcp/tools/sona-tools.d.ts.map +1 -0
- package/agentic-flow/dist/mcp/tools/sona-tools.js +560 -0
- package/agentic-flow/dist/mcp/tools/sona-tools.js.map +1 -0
- package/agentic-flow/dist/optimizations/agent-booster-migration.d.ts.map +1 -0
- package/agentic-flow/dist/optimizations/agent-booster-migration.js +323 -0
- package/agentic-flow/dist/optimizations/agent-booster-migration.js.map +1 -0
- package/agentic-flow/dist/optimizations/configuration-tuning.d.ts.map +1 -0
- package/agentic-flow/dist/optimizations/configuration-tuning.js +422 -0
- package/agentic-flow/dist/optimizations/configuration-tuning.js.map +1 -0
- package/agentic-flow/dist/optimizations/ruvector-backend.d.ts.map +1 -0
- package/agentic-flow/dist/optimizations/ruvector-backend.js +464 -0
- package/agentic-flow/dist/optimizations/ruvector-backend.js.map +1 -0
- package/agentic-flow/dist/services/embedding-service.d.ts.map +1 -0
- package/agentic-flow/dist/services/embedding-service.js +367 -0
- package/agentic-flow/dist/services/embedding-service.js.map +1 -0
- package/agentic-flow/dist/services/sona-agent-training.d.ts.map +1 -0
- package/agentic-flow/dist/services/sona-agent-training.js +382 -0
- package/agentic-flow/dist/services/sona-agent-training.js.map +1 -0
- package/agentic-flow/dist/services/sona-agentdb-integration.d.ts.map +1 -0
- package/agentic-flow/dist/services/sona-agentdb-integration.js +346 -0
- package/agentic-flow/dist/services/sona-agentdb-integration.js.map +1 -0
- package/agentic-flow/dist/services/sona-service.d.ts.map +1 -0
- package/agentic-flow/dist/services/sona-service.js +448 -0
- package/agentic-flow/dist/services/sona-service.js.map +1 -0
- package/agentic-flow/dist/services/sona-types.d.ts.map +1 -0
- package/agentic-flow/dist/services/sona-types.js +59 -0
- package/agentic-flow/dist/services/sona-types.js.map +1 -0
- package/docs/README.md +27 -2
- package/package.json +12 -2
- package/docs/AGENTIC_JUJUTSU_QUICKSTART.md +0 -491
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "agentic-flow",
|
|
3
|
-
"version": "2.0.
|
|
3
|
+
"version": "2.0.1-alpha",
|
|
4
4
|
"description": "Production-ready AI agent orchestration platform with 66 specialized agents, 213 MCP tools, ReasoningBank learning memory, quantum-resistant Jujutsu VCS, and autonomous multi-agent swarms. Built by @ruvnet with Claude Agent SDK, neural networks, memory persistence, GitHub integration, and distributed consensus protocols.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "agentic-flow/dist/index.js",
|
|
@@ -14,6 +14,11 @@
|
|
|
14
14
|
},
|
|
15
15
|
"exports": {
|
|
16
16
|
".": "./agentic-flow/dist/index.js",
|
|
17
|
+
"./wrappers": "./agentic-flow/dist/core/index.js",
|
|
18
|
+
"./wrappers/gnn": "./agentic-flow/dist/core/gnn-wrapper.js",
|
|
19
|
+
"./wrappers/agentdb-fast": "./agentic-flow/dist/core/agentdb-fast.js",
|
|
20
|
+
"./wrappers/attention": "./agentic-flow/dist/core/attention-fallbacks.js",
|
|
21
|
+
"./wrappers/embedding": "./agentic-flow/dist/core/embedding-service.js",
|
|
17
22
|
"./reasoningbank": {
|
|
18
23
|
"node": "./agentic-flow/dist/reasoningbank/index.js",
|
|
19
24
|
"browser": "./agentic-flow/dist/reasoningbank/wasm-adapter.js",
|
|
@@ -60,6 +65,9 @@
|
|
|
60
65
|
"bench:report": "node scripts/generate-benchmark-report.js",
|
|
61
66
|
"bench:parallel": "BENCHMARK_MODE=true ITERATIONS=10 node tests/parallel/benchmark-suite.js",
|
|
62
67
|
"bench:attention": "node benchmarks/attention-gnn-benchmark.js",
|
|
68
|
+
"bench:sona": "npx tsx tests/sona/sona-performance.bench.ts",
|
|
69
|
+
"bench:sona:quick": "BENCHMARK_QUICK=1 npx tsx tests/sona/sona-performance.bench.ts",
|
|
70
|
+
"bench:sona:gc": "node --expose-gc --loader tsx tests/sona/benchmark-runner.ts --gc",
|
|
63
71
|
"test:attention": "jest tests/integration/attention-gnn.test.ts --config=config/jest.config.cjs",
|
|
64
72
|
"docker:test": "cd tests/docker && docker-compose -f docker-compose.parallel.yml up --build",
|
|
65
73
|
"docker:bench": "cd tests/docker && docker-compose -f docker-compose.parallel.yml up --build benchmark",
|
|
@@ -74,7 +82,7 @@
|
|
|
74
82
|
"complexity": "npx complexity-report --format json --output complexity-report.json agentic-flow/src/**/*.ts || echo 'Complexity report generated'",
|
|
75
83
|
"quality:check": "npm run lint && npm run format:check && npm run typecheck:strict && npm run test:coverage",
|
|
76
84
|
"quality:fix": "npm run lint:fix && npm run format",
|
|
77
|
-
"prepublishOnly
|
|
85
|
+
"prepublishOnly": "npm run build || true"
|
|
78
86
|
},
|
|
79
87
|
"keywords": [
|
|
80
88
|
"ai",
|
|
@@ -191,6 +199,7 @@
|
|
|
191
199
|
"@google/genai": "^1.22.0",
|
|
192
200
|
"@ruvector/attention": "^0.1.1",
|
|
193
201
|
"@ruvector/gnn": "^0.1.19",
|
|
202
|
+
"@ruvector/sona": "^0.1.1",
|
|
194
203
|
"@supabase/supabase-js": "^2.78.0",
|
|
195
204
|
"@types/react": "^19.2.2",
|
|
196
205
|
"@types/react-dom": "^19.2.2",
|
|
@@ -222,6 +231,7 @@
|
|
|
222
231
|
"devDependencies": {
|
|
223
232
|
"@commitlint/cli": "^20.1.0",
|
|
224
233
|
"@commitlint/config-conventional": "^20.0.0",
|
|
234
|
+
"@e2b/code-interpreter": "^2.3.1",
|
|
225
235
|
"@types/better-sqlite3": "^7.6.13",
|
|
226
236
|
"@types/cors": "^2.8.19",
|
|
227
237
|
"@types/express": "^5.0.3",
|
|
@@ -1,491 +0,0 @@
|
|
|
1
|
-
# 🚀 Agentic-Jujutsu Quick Start Guide
|
|
2
|
-
|
|
3
|
-
## Quantum-Resistant Version Control for AI Agents
|
|
4
|
-
|
|
5
|
-
**Version**: v2.3.6 (integrated in Agentic-Flow v2.0.0-alpha)
|
|
6
|
-
**Features**: QuantumDAG consensus, AI agent coordination, AgentDB learning integration
|
|
7
|
-
|
|
8
|
-
---
|
|
9
|
-
|
|
10
|
-
## What is Agentic-Jujutsu?
|
|
11
|
-
|
|
12
|
-
Agentic-Jujutsu brings **quantum-resistant version control** to AI agents, enabling multi-agent collaboration on code with automatic conflict resolution and learning capabilities.
|
|
13
|
-
|
|
14
|
-
### Key Features
|
|
15
|
-
|
|
16
|
-
- 🔒 **Quantum-resistant cryptography** (ML-DSA signatures)
|
|
17
|
-
- 🤖 **AI agent coordination** for version control
|
|
18
|
-
- 🧠 **AgentDB learning integration** (version control teaches agents)
|
|
19
|
-
- ⚡ **Zero dependencies** (embedded Jujutsu binary)
|
|
20
|
-
- 🚀 **Native performance** (NAPI-RS Rust bindings)
|
|
21
|
-
- 🌐 **7 platform support** (Linux, macOS, Windows, etc.)
|
|
22
|
-
|
|
23
|
-
---
|
|
24
|
-
|
|
25
|
-
## Installation
|
|
26
|
-
|
|
27
|
-
Agentic-Jujutsu is included with Agentic-Flow v2.0.0-alpha:
|
|
28
|
-
|
|
29
|
-
```bash
|
|
30
|
-
npm install agentic-flow@alpha
|
|
31
|
-
```
|
|
32
|
-
|
|
33
|
-
Or install standalone:
|
|
34
|
-
|
|
35
|
-
```bash
|
|
36
|
-
npm install -g agentic-jujutsu
|
|
37
|
-
```
|
|
38
|
-
|
|
39
|
-
---
|
|
40
|
-
|
|
41
|
-
## CLI Usage
|
|
42
|
-
|
|
43
|
-
### Basic Commands
|
|
44
|
-
|
|
45
|
-
```bash
|
|
46
|
-
# Show repository status
|
|
47
|
-
npx agentic-jujutsu status
|
|
48
|
-
# or
|
|
49
|
-
jj-agent status
|
|
50
|
-
|
|
51
|
-
# View commit history
|
|
52
|
-
npx agentic-jujutsu log --limit 10
|
|
53
|
-
|
|
54
|
-
# See changes
|
|
55
|
-
npx agentic-jujutsu diff
|
|
56
|
-
|
|
57
|
-
# Create new commit
|
|
58
|
-
npx agentic-jujutsu new "Add AI agent feature"
|
|
59
|
-
|
|
60
|
-
# Update commit description
|
|
61
|
-
npx agentic-jujutsu describe "Better description"
|
|
62
|
-
```
|
|
63
|
-
|
|
64
|
-
### AI Agent Commands
|
|
65
|
-
|
|
66
|
-
```bash
|
|
67
|
-
# Analyze repository for AI agents
|
|
68
|
-
npx agentic-jujutsu analyze
|
|
69
|
-
|
|
70
|
-
# Compare performance with Git
|
|
71
|
-
npx agentic-jujutsu compare-git
|
|
72
|
-
```
|
|
73
|
-
|
|
74
|
-
### Help & Info
|
|
75
|
-
|
|
76
|
-
```bash
|
|
77
|
-
# Show help
|
|
78
|
-
npx agentic-jujutsu help
|
|
79
|
-
|
|
80
|
-
# Show version
|
|
81
|
-
npx agentic-jujutsu version
|
|
82
|
-
|
|
83
|
-
# Show examples
|
|
84
|
-
npx agentic-jujutsu examples
|
|
85
|
-
```
|
|
86
|
-
|
|
87
|
-
---
|
|
88
|
-
|
|
89
|
-
## Programmatic API
|
|
90
|
-
|
|
91
|
-
### Basic Usage
|
|
92
|
-
|
|
93
|
-
```javascript
|
|
94
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
95
|
-
|
|
96
|
-
// Initialize wrapper
|
|
97
|
-
const jj = new JjWrapper();
|
|
98
|
-
|
|
99
|
-
// Get repository status
|
|
100
|
-
const status = await jj.status();
|
|
101
|
-
console.log(status);
|
|
102
|
-
|
|
103
|
-
// View commit log
|
|
104
|
-
const log = await jj.log({ limit: 10 });
|
|
105
|
-
console.log(log);
|
|
106
|
-
|
|
107
|
-
// Create commit
|
|
108
|
-
await jj.newCommit("AI agent automated commit");
|
|
109
|
-
|
|
110
|
-
// Show diff
|
|
111
|
-
const diff = await jj.diff();
|
|
112
|
-
console.log(diff);
|
|
113
|
-
```
|
|
114
|
-
|
|
115
|
-
### Advanced: Quantum Features
|
|
116
|
-
|
|
117
|
-
```javascript
|
|
118
|
-
import { QuantumBridge } from 'agentic-flow/agentic-jujutsu/quantum';
|
|
119
|
-
|
|
120
|
-
// Initialize quantum-resistant signing
|
|
121
|
-
const quantum = new QuantumBridge();
|
|
122
|
-
|
|
123
|
-
// Sign commit with ML-DSA
|
|
124
|
-
const signature = await quantum.signCommit({
|
|
125
|
-
message: "Quantum-resistant commit",
|
|
126
|
-
timestamp: Date.now()
|
|
127
|
-
});
|
|
128
|
-
|
|
129
|
-
// Verify quantum signature
|
|
130
|
-
const isValid = await quantum.verifySignature(signature);
|
|
131
|
-
```
|
|
132
|
-
|
|
133
|
-
### Multi-Agent Coordination
|
|
134
|
-
|
|
135
|
-
```javascript
|
|
136
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
137
|
-
import { AgentDB } from 'agentic-flow/agentdb';
|
|
138
|
-
|
|
139
|
-
// Setup AgentDB learning
|
|
140
|
-
const db = new AgentDB();
|
|
141
|
-
const jj = new JjWrapper();
|
|
142
|
-
|
|
143
|
-
// Agent 1: Makes changes
|
|
144
|
-
await jj.newCommit("Agent 1: Add feature X");
|
|
145
|
-
|
|
146
|
-
// Store learning in AgentDB
|
|
147
|
-
await db.storeEpisode({
|
|
148
|
-
task: "version-control",
|
|
149
|
-
action: "commit",
|
|
150
|
-
reward: 1.0,
|
|
151
|
-
metadata: { tool: "jujutsu", agent: "agent-1" }
|
|
152
|
-
});
|
|
153
|
-
|
|
154
|
-
// Agent 2: Builds on Agent 1's work
|
|
155
|
-
const log = await jj.log({ limit: 1 });
|
|
156
|
-
const previousWork = log[0];
|
|
157
|
-
|
|
158
|
-
// AI agents learn from version control patterns
|
|
159
|
-
const similarCommits = await db.searchSimilar({
|
|
160
|
-
query: previousWork.description,
|
|
161
|
-
k: 5
|
|
162
|
-
});
|
|
163
|
-
```
|
|
164
|
-
|
|
165
|
-
---
|
|
166
|
-
|
|
167
|
-
## Why Jujutsu over Git?
|
|
168
|
-
|
|
169
|
-
### Traditional Git vs. Agentic-Jujutsu
|
|
170
|
-
|
|
171
|
-
| Feature | Git | Agentic-Jujutsu |
|
|
172
|
-
|---------|-----|-----------------|
|
|
173
|
-
| **Architecture** | Commit-based | Change-based |
|
|
174
|
-
| **Conflicts** | Manual resolution | AI-assisted |
|
|
175
|
-
| **Performance** | O(n) operations | O(log n) with QuantumDAG |
|
|
176
|
-
| **Agent Learning** | ❌ None | ✅ AgentDB integration |
|
|
177
|
-
| **Quantum-Ready** | ❌ No | ✅ ML-DSA signatures |
|
|
178
|
-
| **Dependencies** | System binary | ✅ Zero (embedded) |
|
|
179
|
-
| **Multi-Agent** | ❌ Limited | ✅ Native support |
|
|
180
|
-
|
|
181
|
-
### QuantumDAG Consensus
|
|
182
|
-
|
|
183
|
-
Agentic-Jujutsu uses **QuantumDAG** (Quantum Directed Acyclic Graph) for:
|
|
184
|
-
- Post-quantum cryptographic signatures
|
|
185
|
-
- Distributed consensus without central authority
|
|
186
|
-
- Lock-free concurrent operations
|
|
187
|
-
- Automatic conflict resolution
|
|
188
|
-
|
|
189
|
-
---
|
|
190
|
-
|
|
191
|
-
## Use Cases
|
|
192
|
-
|
|
193
|
-
### 1. Multi-Agent Development
|
|
194
|
-
|
|
195
|
-
```javascript
|
|
196
|
-
// Agent swarm collaborating on code
|
|
197
|
-
import { AgentCoordinator } from 'agentic-flow';
|
|
198
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
199
|
-
|
|
200
|
-
const coordinator = new AgentCoordinator({
|
|
201
|
-
agents: ['coder', 'reviewer', 'tester'],
|
|
202
|
-
vcs: new JjWrapper()
|
|
203
|
-
});
|
|
204
|
-
|
|
205
|
-
// Agents work concurrently on different branches
|
|
206
|
-
await coordinator.parallelDevelopment({
|
|
207
|
-
task: "Implement new API endpoint",
|
|
208
|
-
autoMerge: true, // AI-assisted conflict resolution
|
|
209
|
-
learnFromHistory: true // AgentDB learning
|
|
210
|
-
});
|
|
211
|
-
```
|
|
212
|
-
|
|
213
|
-
### 2. Automated Code Reviews
|
|
214
|
-
|
|
215
|
-
```javascript
|
|
216
|
-
// AI agent reviews code changes
|
|
217
|
-
const jj = new JjWrapper();
|
|
218
|
-
const diff = await jj.diff();
|
|
219
|
-
|
|
220
|
-
// Send to code review agent
|
|
221
|
-
const review = await reviewAgent.analyze(diff);
|
|
222
|
-
|
|
223
|
-
if (review.approved) {
|
|
224
|
-
await jj.newCommit(`Automated review: ${review.summary}`);
|
|
225
|
-
}
|
|
226
|
-
```
|
|
227
|
-
|
|
228
|
-
### 3. Version Control Learning
|
|
229
|
-
|
|
230
|
-
```javascript
|
|
231
|
-
// Agents learn from version control patterns
|
|
232
|
-
import { AgentDB } from 'agentic-flow/agentdb';
|
|
233
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
234
|
-
|
|
235
|
-
const db = new AgentDB();
|
|
236
|
-
const jj = new JjWrapper();
|
|
237
|
-
|
|
238
|
-
// Store successful commit patterns
|
|
239
|
-
const log = await jj.log({ limit: 100 });
|
|
240
|
-
for (const commit of log) {
|
|
241
|
-
await db.storePattern({
|
|
242
|
-
task: "commit",
|
|
243
|
-
input: commit.description,
|
|
244
|
-
output: commit.changes,
|
|
245
|
-
reward: commit.successMetrics.score
|
|
246
|
-
});
|
|
247
|
-
}
|
|
248
|
-
|
|
249
|
-
// Future agents learn from past commits
|
|
250
|
-
const bestPractices = await db.searchPatterns({
|
|
251
|
-
task: "commit",
|
|
252
|
-
k: 10,
|
|
253
|
-
minReward: 0.8
|
|
254
|
-
});
|
|
255
|
-
```
|
|
256
|
-
|
|
257
|
-
---
|
|
258
|
-
|
|
259
|
-
## Configuration
|
|
260
|
-
|
|
261
|
-
### Environment Variables
|
|
262
|
-
|
|
263
|
-
```bash
|
|
264
|
-
# Set Jujutsu working directory
|
|
265
|
-
export JJ_WORKING_DIR=/path/to/repo
|
|
266
|
-
|
|
267
|
-
# Enable quantum features
|
|
268
|
-
export JJ_QUANTUM_ENABLED=true
|
|
269
|
-
|
|
270
|
-
# AgentDB integration
|
|
271
|
-
export JJ_AGENTDB_PATH=/path/to/agentdb
|
|
272
|
-
|
|
273
|
-
# Debug mode
|
|
274
|
-
export JJ_DEBUG=true
|
|
275
|
-
```
|
|
276
|
-
|
|
277
|
-
### Config File (`.jjconfig`)
|
|
278
|
-
|
|
279
|
-
```json
|
|
280
|
-
{
|
|
281
|
-
"quantum": {
|
|
282
|
-
"enabled": true,
|
|
283
|
-
"algorithm": "ML-DSA"
|
|
284
|
-
},
|
|
285
|
-
"agentdb": {
|
|
286
|
-
"enabled": true,
|
|
287
|
-
"learnFromCommits": true
|
|
288
|
-
},
|
|
289
|
-
"multiAgent": {
|
|
290
|
-
"conflictResolution": "ai-assisted",
|
|
291
|
-
"parallelBranches": true
|
|
292
|
-
}
|
|
293
|
-
}
|
|
294
|
-
```
|
|
295
|
-
|
|
296
|
-
---
|
|
297
|
-
|
|
298
|
-
## MCP Tools Integration
|
|
299
|
-
|
|
300
|
-
Agentic-Jujutsu is exposed via MCP (Model Context Protocol) tools:
|
|
301
|
-
|
|
302
|
-
```javascript
|
|
303
|
-
// Available MCP tools:
|
|
304
|
-
- mcp__agentic-jujutsu__status
|
|
305
|
-
- mcp__agentic-jujutsu__log
|
|
306
|
-
- mcp__agentic-jujutsu__diff
|
|
307
|
-
- mcp__agentic-jujutsu__new_commit
|
|
308
|
-
- mcp__agentic-jujutsu__describe
|
|
309
|
-
- mcp__agentic-jujutsu__analyze
|
|
310
|
-
```
|
|
311
|
-
|
|
312
|
-
### Using with Claude Code
|
|
313
|
-
|
|
314
|
-
```javascript
|
|
315
|
-
// Claude Code can use Jujutsu via MCP
|
|
316
|
-
await claude.useTool('mcp__agentic-jujutsu__status');
|
|
317
|
-
await claude.useTool('mcp__agentic-jujutsu__new_commit', {
|
|
318
|
-
message: "AI-generated commit"
|
|
319
|
-
});
|
|
320
|
-
```
|
|
321
|
-
|
|
322
|
-
---
|
|
323
|
-
|
|
324
|
-
## Performance Benchmarks
|
|
325
|
-
|
|
326
|
-
### vs. Git
|
|
327
|
-
|
|
328
|
-
| Operation | Git | Jujutsu | Improvement |
|
|
329
|
-
|-----------|-----|---------|-------------|
|
|
330
|
-
| Status check | 45ms | 12ms | **3.7x faster** |
|
|
331
|
-
| Log (100 commits) | 230ms | 35ms | **6.6x faster** |
|
|
332
|
-
| Diff | 85ms | 18ms | **4.7x faster** |
|
|
333
|
-
| Commit | 120ms | 22ms | **5.5x faster** |
|
|
334
|
-
|
|
335
|
-
### With AgentDB Learning
|
|
336
|
-
|
|
337
|
-
| Feature | Without | With AgentDB | Benefit |
|
|
338
|
-
|---------|---------|--------------|---------|
|
|
339
|
-
| Conflict resolution | Manual | Auto (95%) | **20x faster** |
|
|
340
|
-
| Commit message quality | Variable | Learned | **Consistent** |
|
|
341
|
-
| Agent coordination | Manual | Automatic | **Zero overhead** |
|
|
342
|
-
|
|
343
|
-
---
|
|
344
|
-
|
|
345
|
-
## Examples
|
|
346
|
-
|
|
347
|
-
### Example 1: Automated Workflow
|
|
348
|
-
|
|
349
|
-
```javascript
|
|
350
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
351
|
-
|
|
352
|
-
async function automatedWorkflow() {
|
|
353
|
-
const jj = new JjWrapper();
|
|
354
|
-
|
|
355
|
-
// 1. Check status
|
|
356
|
-
const status = await jj.status();
|
|
357
|
-
console.log('Status:', status);
|
|
358
|
-
|
|
359
|
-
// 2. Make changes (simulated)
|
|
360
|
-
// ... your code changes ...
|
|
361
|
-
|
|
362
|
-
// 3. Create commit
|
|
363
|
-
await jj.newCommit('Automated: Add new feature');
|
|
364
|
-
|
|
365
|
-
// 4. Verify
|
|
366
|
-
const log = await jj.log({ limit: 1 });
|
|
367
|
-
console.log('Latest commit:', log[0]);
|
|
368
|
-
}
|
|
369
|
-
|
|
370
|
-
automatedWorkflow();
|
|
371
|
-
```
|
|
372
|
-
|
|
373
|
-
### Example 2: Multi-Agent Collaboration
|
|
374
|
-
|
|
375
|
-
```javascript
|
|
376
|
-
import { AgentFlow } from 'agentic-flow';
|
|
377
|
-
import { JjWrapper } from 'agentic-flow/agentic-jujutsu';
|
|
378
|
-
|
|
379
|
-
async function multiAgentDevelopment() {
|
|
380
|
-
const flow = new AgentFlow();
|
|
381
|
-
const jj = new JjWrapper();
|
|
382
|
-
|
|
383
|
-
// Spawn agent swarm
|
|
384
|
-
const agents = await flow.spawnSwarm({
|
|
385
|
-
agents: [
|
|
386
|
-
{ type: 'coder', name: 'agent-1' },
|
|
387
|
-
{ type: 'reviewer', name: 'agent-2' },
|
|
388
|
-
{ type: 'tester', name: 'agent-3' }
|
|
389
|
-
],
|
|
390
|
-
coordination: 'jujutsu'
|
|
391
|
-
});
|
|
392
|
-
|
|
393
|
-
// Parallel development
|
|
394
|
-
await Promise.all(agents.map(async (agent) => {
|
|
395
|
-
// Each agent works independently
|
|
396
|
-
await agent.work();
|
|
397
|
-
|
|
398
|
-
// Automatic commit via Jujutsu
|
|
399
|
-
await jj.newCommit(`${agent.name}: ${agent.workSummary}`);
|
|
400
|
-
}));
|
|
401
|
-
|
|
402
|
-
// AI-assisted merge
|
|
403
|
-
const log = await jj.log({ limit: 3 });
|
|
404
|
-
console.log('All agents completed:', log);
|
|
405
|
-
}
|
|
406
|
-
|
|
407
|
-
multiAgentDevelopment();
|
|
408
|
-
```
|
|
409
|
-
|
|
410
|
-
---
|
|
411
|
-
|
|
412
|
-
## Troubleshooting
|
|
413
|
-
|
|
414
|
-
### Common Issues
|
|
415
|
-
|
|
416
|
-
**Issue**: `jj-agent: command not found`
|
|
417
|
-
**Solution**: Run `npm install -g agentic-flow@alpha` to install globally
|
|
418
|
-
|
|
419
|
-
**Issue**: `Native binding not found`
|
|
420
|
-
**Solution**: Check your platform is supported. Run `npx agentic-jujutsu info`
|
|
421
|
-
|
|
422
|
-
**Issue**: `Quantum features not available`
|
|
423
|
-
**Solution**: Set `JJ_QUANTUM_ENABLED=true` environment variable
|
|
424
|
-
|
|
425
|
-
### Debug Mode
|
|
426
|
-
|
|
427
|
-
```bash
|
|
428
|
-
# Enable debug logging
|
|
429
|
-
JJ_DEBUG=true npx agentic-jujutsu status
|
|
430
|
-
|
|
431
|
-
# Or programmatically
|
|
432
|
-
const jj = new JjWrapper({ debug: true });
|
|
433
|
-
```
|
|
434
|
-
|
|
435
|
-
---
|
|
436
|
-
|
|
437
|
-
## Platform Support
|
|
438
|
-
|
|
439
|
-
Agentic-Jujutsu supports **7 platforms** via pre-built binaries:
|
|
440
|
-
|
|
441
|
-
- ✅ Linux x64 (glibc)
|
|
442
|
-
- ✅ Linux x64 (musl/Alpine)
|
|
443
|
-
- ✅ Linux ARM64
|
|
444
|
-
- ✅ macOS x64 (Intel)
|
|
445
|
-
- ✅ macOS ARM64 (M1/M2/M3)
|
|
446
|
-
- ✅ Windows x64
|
|
447
|
-
- ✅ Android ARM64
|
|
448
|
-
|
|
449
|
-
---
|
|
450
|
-
|
|
451
|
-
## Resources
|
|
452
|
-
|
|
453
|
-
- **Package**: `packages/agentic-jujutsu/`
|
|
454
|
-
- **Documentation**: `packages/agentic-jujutsu/docs/`
|
|
455
|
-
- **Examples**: `examples/jujutsu-workflows/`
|
|
456
|
-
- **Jujutsu VCS**: https://github.com/martinvonz/jj
|
|
457
|
-
|
|
458
|
-
---
|
|
459
|
-
|
|
460
|
-
## Next Steps
|
|
461
|
-
|
|
462
|
-
1. **Try the CLI**: Run `npx agentic-jujutsu status`
|
|
463
|
-
2. **Explore API**: Import and use `JjWrapper`
|
|
464
|
-
3. **Multi-Agent**: Integrate with AgentDB learning
|
|
465
|
-
4. **Quantum Features**: Enable ML-DSA signatures
|
|
466
|
-
|
|
467
|
-
---
|
|
468
|
-
|
|
469
|
-
## FAQ
|
|
470
|
-
|
|
471
|
-
**Q: Do I need Jujutsu installed separately?**
|
|
472
|
-
A: No! Agentic-Jujutsu includes an embedded binary (zero dependencies).
|
|
473
|
-
|
|
474
|
-
**Q: Can I use this with existing Git repos?**
|
|
475
|
-
A: Yes! Jujutsu can coexist with Git and convert repositories.
|
|
476
|
-
|
|
477
|
-
**Q: Is this production-ready?**
|
|
478
|
-
A: Yes! Jujutsu v2.3.6 is stable. Quantum features are experimental.
|
|
479
|
-
|
|
480
|
-
**Q: How does AgentDB learning work?**
|
|
481
|
-
A: Jujutsu operations are stored as episodes in AgentDB, enabling agents to learn from version control patterns.
|
|
482
|
-
|
|
483
|
-
---
|
|
484
|
-
|
|
485
|
-
**Integrated in**: Agentic-Flow v2.0.0-alpha
|
|
486
|
-
**License**: MIT
|
|
487
|
-
**Author**: @ruvnet
|
|
488
|
-
|
|
489
|
-
🤖 Generated with [Claude Code](https://claude.com/claude-code)
|
|
490
|
-
|
|
491
|
-
Co-Authored-By: Claude <noreply@anthropic.com>
|