FangCode enables a single developer to perform the work of an entire engineering team. It provides 31 specialized agents, multi-agent orchestration with team-based execution, consensus voting, and seamless MiMoCode integration through slash commands.
git clone https://github.com/aydocs/FangCode.git
cd FangCode
npm install
node bin/fc.js agents
npm install -g .
fc agents
Copy the skills to your MiMoCode config directory:
# Windows
copy skills\* %USERPROFILE%\.config\mimocode\skills\
# macOS / Linux
cp -r skills/* ~/.config/mimocode/skills/
Then add FangCode to ~/.config/mimocode/mimocode.json:
{
"skills": {
"fangcode": "C:\\path\\to\\FangCode"
}
}Restart MiMoCode. All skills load automatically.
| Module |
Purpose |
| Agent Registry |
Store, index, and query agents by category, tag, or full-text search |
| Agent Manager |
Install, remove, update, run, export, and import agents |
| Agent Loader |
Parse YAML and JSON agent definitions with error handling |
| Agent Marketplace |
Discover, rate, and track agent downloads |
| Agent Templates |
Generate new agents from predefined templates |
| Agent Discovery |
Scan local directories, remote registries, and MCP servers |
| Agent Profiles |
Persist user preferences, favorites, and usage history |
| Agent Routing |
Match requests to the most suitable agent using keyword and skill analysis |
| Agent Orchestration |
Chain, parallel, fan-out, and fan-in execution patterns |
| Agent Workflows |
Define and execute DAG-based multi-step pipelines |
| MCP Integration |
Connect to MCP servers, invoke tools, and manage resources |
| Multi-Agent System |
Create teams, send messages, and reach consensus through voting |
| Custom Commands |
14 slash commands for direct MiMoCode integration |
| Command |
Description |
/agents |
List all registered agents |
/list-agents |
List agents with detailed information |
/install-agent |
Install an agent from a file or URL |
/remove-agent |
Remove an installed agent |
/update-agent |
Update an agent to its latest version |
/run-agent |
Execute an agent with a given task |
/search-agent |
Search agents by keyword |
/export-agent |
Export an agent definition to a file |
/import-agent |
Import an agent from a file |
/team-create |
Create a multi-agent team |
/team-execute |
Execute a task using a team of agents |
/team-list |
List all active teams |
/team-agents |
List agents participating in multi-agent workflows |
/consensus |
Request a team vote on a proposal |
| Agent |
Category |
Expertise |
web-developer |
web |
HTML, CSS, JavaScript, TypeScript, responsive design, accessibility, SEO |
frontend-developer |
frontend |
React, Vue, Angular, Svelte, component architecture, state management |
full-stack-developer |
fullstack |
Frontend, backend, databases, DevOps, testing, authentication |
| Agent |
Category |
Expertise |
backend-developer |
backend |
REST, GraphQL, gRPC, databases, authentication, microservices, caching |
database-engineer |
database |
PostgreSQL, MySQL, MongoDB, Redis, data modeling, query optimization |
architect |
fullstack |
System design, DDD, CQRS, event-driven architecture, scalability |
| Agent |
Category |
Expertise |
mobile-developer |
mobile |
Flutter, React Native, Swift, Kotlin, Dart, mobile architecture |
flutter-developer |
mobile |
Flutter, Dart, widgets, state management, animations, platform channels |
android-developer |
mobile |
Kotlin, Jetpack Compose, Android architecture, Gradle, Room, Coroutines |
ios-developer |
mobile |
Swift, SwiftUI, UIKit, Combine, Core Data, Xcode |
| Agent |
Category |
Expertise |
devops-engineer |
devops |
Docker, Kubernetes, Terraform, Ansible, CI/CD, monitoring, GitHub Actions |
cloud-architect |
cloud |
AWS, GCP, Azure, serverless, microservices, IAM, networking |
kubernetes-engineer |
devops |
Kubernetes, Helm, Kustomize, Istio, container security, RBAC, observability |
| Agent |
Category |
Expertise |
security-auditor |
security |
OWASP, SAST, DAST, dependency scanning, compliance |
secure-code-reviewer |
security |
Code review, static analysis, secure coding, cryptography |
threat-modeler |
security |
STRIDE, PASTA, attack trees, risk analysis, MITRE ATT&CK |
soc-analyst |
security |
SIEM, Sigma rules, threat hunting, forensics |
incident-responder |
security |
Incident response, digital forensics, breach management |
malware-analyst |
security |
Malware analysis, YARA rules, threat intelligence |
reverse-engineer |
security |
Binary analysis, Ghidra, IDA, x86, ARM, RISC-V |
| Agent |
Category |
Expertise |
ai-engineer |
ai |
PyTorch, TensorFlow, MLOps, model deployment, NLP |
prompt-engineer |
ai |
Prompt engineering, chain-of-thought, few-shot learning, LLM optimization |
agent-builder |
ai |
Agent design, tool use, memory systems, multi-agent orchestration |
mcp-engineer |
ai |
MCP server development, tool schemas, JSON-RPC, transports |
| Agent |
Category |
Expertise |
rust-developer |
language |
Ownership, lifetimes, async Rust, Cargo, WASM |
go-developer |
language |
Goroutines, channels, interfaces, Cobra, gRPC |
python-developer |
language |
Django, Flask, FastAPI, type hints, asyncio, packaging |
nodejs-developer |
language |
Express, Fastify, NestJS, TypeScript, streams |
| Agent |
Category |
Expertise |
linux-engineer |
linux |
Shell scripting, systemd, networking, security hardening |
startup-cto |
leadership |
Architecture, team building, technical debt, hiring, agile |
FangCode includes a full multi-agent system with team-based execution.
const { FangCode } = require('./src');
const fc = new FangCode();
await fc.init();
fc.multiagent.createTeam('backend-team', [
'go-developer',
'python-developer',
'database-engineer'
], { strategy: 'parallel' });
const results = await fc.multiagent.teamExecute('backend-team', 'Build a microservice API');
| Strategy |
Description |
sequential |
Agents execute one after another, passing context between steps |
parallel |
All agents execute simultaneously for maximum throughput |
leader-follower |
A designated leader makes decisions; followers implement |
debate |
Agents argue different perspectives; the leader synthesizes a final answer |
const consensus = await fc.multiagent.requestConsensus('backend-team',
'Should we use PostgreSQL instead of MongoDB?'
);
console.log(consensus.approved ? 'APPROVED' : 'REJECTED');
const { FangCode } = require('./src');
const fc = new FangCode();
await fc.init();
// List agents by category
const securityAgents = fc.manager.list({ category: 'security' });
// Search agents
const pythonAgents = fc.manager.search('python');
// Run an agent
const config = await fc.manager.run('devops-engineer', {
task: 'Set up monitoring for production'
});
// Route a request to the best agent
const match = fc.route('I need help with a React dashboard');
// Execute a workflow
await fc.executeWorkflow({
name: 'fullstack-build',
steps: [
{ id: 'design', agent: 'architect', task: 'System design' },
{ id: 'frontend', agent: 'frontend-developer', task: 'Build UI', depends_on: ['design'] },
{ id: 'backend', agent: 'backend-developer', task: 'Build API', depends_on: ['design'] },
{ id: 'deploy', agent: 'devops-engineer', task: 'Deploy', depends_on: ['frontend', 'backend'] },
]
});
FangCode ships with production-grade design skills from leading open-source projects.
fangcode/
├── agents/ # 31 agent definitions (YAML)
├── bin/
│ └── fc.js # CLI entry point
├── src/
│ ├── index.js # Main framework export
│ ├── registry/ # Agent storage and indexing
│ ├── manager/ # Install, remove, run, export, import
│ ├── loader/ # YAML/JSON parsing
│ ├── marketplace/ # Discovery, ratings, downloads
│ ├── templates/ # Agent creation templates
│ ├── discovery/ # Local, remote, and MCP scanning
│ ├── profiles/ # User preferences and history
│ ├── routing/ # Request-to-agent matching
│ ├── orchestration/ # Chain, parallel, fan-out, fan-in
│ ├── workflows/ # DAG-based pipeline engine
│ ├── mcp/ # MCP server connections
│ ├── multiagent/ # Teams, messaging, consensus
│ ├── commands/ # 14 slash commands
│ └── schemas/ # YAML and JSON agent schemas
├── skills/ # Design engineering skills
├── templates/ # Agent creation templates
├── docs/
├── examples/
└── tests/
FangCode stores its configuration at ~/.fangcode/config.yaml:
version: "1.0.0"
log_level: info
agents_dir: ./agents
mcp:
enabled: true
servers: {}
routing:
auto_route: true
fallback_agent: full-stack-developer
multiagent:
max_concurrent: 8
timeout_ms: 300000
protocol: async
- Fork the repository
- Create a feature branch
- Commit your changes
- Open a pull request
To add a new agent, place a YAML file in the agents/ directory following the schema in src/schemas/agent-schema.yaml.
To add a new skill, create a SKILL.md file in the skills/ directory following the MiMoCode skill format.
MIT © FangCode Community
