krenk 0.1.0
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/LICENSE +21 -0
- package/README.md +172 -0
- package/bin/krenk.js +2 -0
- package/dist/index.js +3880 -0
- package/dist/index.js.map +1 -0
- package/package.json +59 -0
- package/public/img/main.png +0 -0
package/LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Krishna Kant Singh
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
package/README.md
ADDED
|
@@ -0,0 +1,172 @@
|
|
|
1
|
+
<p align="center">
|
|
2
|
+
<img src="public/img/main.png" alt="Krenk CLI" width="700" />
|
|
3
|
+
</p>
|
|
4
|
+
|
|
5
|
+
# Krenk
|
|
6
|
+
|
|
7
|
+
**Like Claude Code, but with a team.**
|
|
8
|
+
|
|
9
|
+
Krenk is a multi-agent software engineering CLI that orchestrates an entire team of specialized AI agents to build your project -- from planning to deployment. Instead of one AI doing everything, Krenk assigns the right job to the right agent, just like a real engineering team.
|
|
10
|
+
|
|
11
|
+
---
|
|
12
|
+
|
|
13
|
+
## Why Krenk over plain Claude Code?
|
|
14
|
+
|
|
15
|
+
Claude Code is powerful, but it's one agent doing everything: planning, coding, testing, reviewing, documenting. That's like asking a single developer to be the architect, the frontend engineer, the QA lead, the security auditor, and the tech writer -- all at once, in one context window.
|
|
16
|
+
|
|
17
|
+
**Krenk fixes this.**
|
|
18
|
+
|
|
19
|
+
| | Claude Code | Krenk |
|
|
20
|
+
|---|---|---|
|
|
21
|
+
| Agents | 1 | Up to 11 specialized agents |
|
|
22
|
+
| Planning | Inline, ad-hoc | Dedicated Strategist creates a master plan with per-agent assignments |
|
|
23
|
+
| Architecture | Mixed with implementation | Dedicated Architect sets up skeleton before Builder writes features |
|
|
24
|
+
| Code review | You review manually | Sentinel agent reviews code automatically for bugs, security, and quality |
|
|
25
|
+
| Testing | You write tests or ask | Guardian agent writes and runs comprehensive test suites |
|
|
26
|
+
| Security | Hope for the best | Shield agent performs a dedicated security audit |
|
|
27
|
+
| Coordination | Single context window | Master Brain coordinates agents via shared memory |
|
|
28
|
+
| Supervision | None | Process Supervisor monitors memory, CPU, and kills hung agents |
|
|
29
|
+
| Parallelism | Sequential | Parallel builders for independent modules |
|
|
30
|
+
|
|
31
|
+
**The result:** Higher quality output, faster execution for large projects, and specialized attention to each concern -- planning, design, architecture, implementation, testing, review, security, documentation, and deployment all get dedicated focus.
|
|
32
|
+
|
|
33
|
+
---
|
|
34
|
+
|
|
35
|
+
## How it works
|
|
36
|
+
|
|
37
|
+
1. **You describe what you want to build** -- plain English, typos and all
|
|
38
|
+
2. **Krenk refines your prompt** -- using Claude to extract requirements, tech stack, and structure
|
|
39
|
+
3. **The Strategist creates a master plan** -- breaking work into tasks and assigning them to agents
|
|
40
|
+
4. **Agents execute in sequence** -- each one focused on its specialty, building on the previous agent's output
|
|
41
|
+
5. **The Master Brain coordinates** -- sharing context, reviewing output, ordering redos if quality is low
|
|
42
|
+
6. **The Supervisor watches everything** -- monitoring memory, CPU, and killing hung processes
|
|
43
|
+
|
|
44
|
+
### The Team
|
|
45
|
+
|
|
46
|
+
| Agent | Role | What it does |
|
|
47
|
+
|---|---|---|
|
|
48
|
+
| **Analyst** | Business analysis | User stories, acceptance criteria, priorities |
|
|
49
|
+
| **Strategist** | Planning | Master plan with per-agent task assignments |
|
|
50
|
+
| **Pixel** | UI/UX design | Component hierarchy, layouts, styling specs |
|
|
51
|
+
| **Blueprint** | Architecture | System design, project skeleton, data models |
|
|
52
|
+
| **Builder** | Implementation | Writes the actual production code |
|
|
53
|
+
| **QA Lead** | Test planning | Test strategy, test cases, coverage matrix |
|
|
54
|
+
| **Guardian** | Testing | Writes and runs unit/integration tests |
|
|
55
|
+
| **Sentinel** | Code review | Reviews for bugs, performance, and style |
|
|
56
|
+
| **Shield** | Security audit | OWASP checks, dependency audit, auth review |
|
|
57
|
+
| **Scribe** | Documentation | README, API docs, inline comments |
|
|
58
|
+
| **DevOps** | Deployment | CI/CD, Docker, deployment configs |
|
|
59
|
+
|
|
60
|
+
Not every project needs all 11. Krenk lets you pick your team:
|
|
61
|
+
|
|
62
|
+
- **Full Team** -- all 11 agents
|
|
63
|
+
- **Engineering** -- plan, architect, code, test, review
|
|
64
|
+
- **QA Focused** -- plan, code, QA, test, review
|
|
65
|
+
- **Startup MVP** -- plan, architect, code, review, docs
|
|
66
|
+
- **Quick Build** -- plan and code only
|
|
67
|
+
- **Custom** -- pick your own agents
|
|
68
|
+
|
|
69
|
+
---
|
|
70
|
+
|
|
71
|
+
## Installation
|
|
72
|
+
|
|
73
|
+
### Prerequisites
|
|
74
|
+
|
|
75
|
+
- Node.js >= 20
|
|
76
|
+
- [Claude Code CLI](https://docs.anthropic.com/en/docs/claude-code) installed and authenticated
|
|
77
|
+
|
|
78
|
+
### Install from source
|
|
79
|
+
|
|
80
|
+
```bash
|
|
81
|
+
git clone https://github.com/yourusername/krenk.git
|
|
82
|
+
cd krenk
|
|
83
|
+
npm install
|
|
84
|
+
npm run build
|
|
85
|
+
npm link
|
|
86
|
+
```
|
|
87
|
+
|
|
88
|
+
### Run
|
|
89
|
+
|
|
90
|
+
```bash
|
|
91
|
+
krenk
|
|
92
|
+
```
|
|
93
|
+
|
|
94
|
+
That's it. The interactive UI walks you through everything.
|
|
95
|
+
|
|
96
|
+
---
|
|
97
|
+
|
|
98
|
+
## Architecture
|
|
99
|
+
|
|
100
|
+
```
|
|
101
|
+
krenk
|
|
102
|
+
|-- src/
|
|
103
|
+
| |-- orchestrator/
|
|
104
|
+
| | |-- engine.ts # Main pipeline: runs agents through stages
|
|
105
|
+
| | |-- brain.ts # Master Brain: coordinates agents, reviews, redos
|
|
106
|
+
| | |-- memory.ts # Shared memory: .krenk/shared/ files all agents read/write
|
|
107
|
+
| | |-- supervisor.ts # Process watchdog: memory, CPU, timeout monitoring
|
|
108
|
+
| | |-- plan-parser.ts # Parses strategist output into per-agent assignments
|
|
109
|
+
| | |-- scheduler.ts # Parallel execution with concurrency control
|
|
110
|
+
| | |-- context.ts # Accumulates output from previous agents
|
|
111
|
+
| | |-- workflow.ts # Stage definitions and ordering
|
|
112
|
+
| |-- agents/
|
|
113
|
+
| | |-- roles.ts # Agent definitions: prompts, tools, boundaries
|
|
114
|
+
| | |-- spawner.ts # Spawns Claude CLI processes with stream-json
|
|
115
|
+
| | |-- registry.ts # Tracks active agents, PIDs, output
|
|
116
|
+
| |-- ui/
|
|
117
|
+
| | |-- interactive.ts # Interactive CLI: team selection, prompt, refinement
|
|
118
|
+
| | |-- renderer.ts # Real-time spinner with stream-json activity
|
|
119
|
+
| | |-- theme.ts # Colors and gradients
|
|
120
|
+
| |-- config/
|
|
121
|
+
| | |-- defaults.ts # Default config: max turns, parallel agents
|
|
122
|
+
| | |-- loader.ts # Config file loader (cosmiconfig)
|
|
123
|
+
| |-- utils/
|
|
124
|
+
| |-- logger.ts # Logging
|
|
125
|
+
| |-- process.ts # Graceful shutdown
|
|
126
|
+
```
|
|
127
|
+
|
|
128
|
+
### Key design decisions
|
|
129
|
+
|
|
130
|
+
**Stream-JSON output** -- Agents use `--output-format stream-json` so the CLI shows real-time activity (which file is being read, which tool is being called) instead of a frozen spinner.
|
|
131
|
+
|
|
132
|
+
**Tool restrictions** -- Read-only agents (Strategist, Analyst, Sentinel) have `--disallowedTools Write,Edit,Bash` enforced at the CLI level so they cannot accidentally write code.
|
|
133
|
+
|
|
134
|
+
**Shared memory** -- All agents can read `.krenk/shared/` markdown files. The Master Brain writes directives to `brain.md`, agents write their progress. This lets later agents build on earlier agents' decisions without passing massive context strings.
|
|
135
|
+
|
|
136
|
+
**Process supervision** -- Every spawned Claude process is tracked. If an agent exceeds 512MB memory, runs longer than 10 minutes, or produces no output for 5 minutes, the Supervisor kills it and the Brain can retry.
|
|
137
|
+
|
|
138
|
+
---
|
|
139
|
+
|
|
140
|
+
## Configuration
|
|
141
|
+
|
|
142
|
+
Create a `.krenkrc.json`, `.krenkrc.yml`, or `krenk.config.js` in your project root:
|
|
143
|
+
|
|
144
|
+
```json
|
|
145
|
+
{
|
|
146
|
+
"maxParallelAgents": 3,
|
|
147
|
+
"agents": {
|
|
148
|
+
"strategist": { "maxTurns": 30 },
|
|
149
|
+
"architect": { "maxTurns": 50 },
|
|
150
|
+
"builder": { "maxTurns": 100 },
|
|
151
|
+
"guardian": { "maxTurns": 50 }
|
|
152
|
+
}
|
|
153
|
+
}
|
|
154
|
+
```
|
|
155
|
+
|
|
156
|
+
---
|
|
157
|
+
|
|
158
|
+
## Modes
|
|
159
|
+
|
|
160
|
+
**Autonomous** -- agents run without asking. Fastest path from prompt to project.
|
|
161
|
+
|
|
162
|
+
**Supervised** -- you approve each agent before it runs. You see what it will do and can skip or abort.
|
|
163
|
+
|
|
164
|
+
---
|
|
165
|
+
|
|
166
|
+
## License
|
|
167
|
+
|
|
168
|
+
MIT
|
|
169
|
+
|
|
170
|
+
---
|
|
171
|
+
|
|
172
|
+
Built by Krishna with mass amount of coffee.
|
package/bin/krenk.js
ADDED