A powerful, extensible framework for building AI agents with comprehensive tool support, multi-agent orchestration, and advanced web automation capabilities.

ManusUse empowers developers to create sophisticated AI agents that can:

• Execute code in secure Docker sandboxes
• Automate web browsing and data extraction
• Analyze data and generate visualizations
• Coordinate multiple specialized agents for complex tasks
• Integrate with various LLM providers seamlessly
• Perform deep vulnerability intelligence analysis

Built on Strands SDK and integrated with browser-use, ManusUse provides a production-ready foundation for AI agent development.

• Installation
• Quick Start
• CLI Reference
• Configuration
• Python API
• Security & Vulnerability Intelligence
• Development

# Basic installation
pip install manus-use

# With browser automation support
pip install manus-use[browser]
playwright install chromium

# Full installation with all optional features
pip install manus-use[browser,search,visualization]

manus-use init

The interactive wizard creates ~/.manus-use/config.toml with your LLM provider credentials.

manus-use doctor

Verifies installed packages, configuration, and API key accessibility.

# Single-shot (non-interactive)
manus-use "Write a Python script that fetches the current Bitcoin price"

# Interactive REPL
manus-use

ManusUse ships a single manus-use entry point with several subcommands.

# Single-shot task (prints result, then exits)
manus-use "Create a factorial function in Python"

# Use a specific agent type
manus-use --agent browser "Find the top 5 trending GitHub repos today"

# Force multi-agent orchestration
manus-use --mode multi "Research quantum computing and create a presentation"

# Save output to a file
manus-use --output result.txt "Summarise the latest AI news"

# JSON output for piping into other tools
manus-use --format json "List the first 10 prime numbers" | jq .result

# Interactive REPL (omit the task argument)
manus-use
manus-use --mode multi

Options:

manus-use init                       # write to ~/.manus-use/config.toml
manus-use init --output ./my.toml   # write to a custom location
manus-use init --force               # overwrite without prompting

manus-use doctor
manus-use doctor --config ./custom.toml

Checks Python packages, config file validity, and whether API keys are accessible.

# Deep CVE analysis (NVD · CISA KEV · OTX · PoC search · CWE · threat feeds)
manus-use analyze CVE-2025-6554

# With Docker-based exploit verification
manus-use analyze CVE-2024-3094 --verify

# Machine-readable output
manus-use analyze CVE-2025-6554 --output json

# Generate a Lark document report
manus-use analyze CVE-2025-6554 --output lark

Options:

manus-use history                        # last 20 runs
manus-use history --limit 50             # last 50 runs
manus-use history --grep "bitcoin"       # filter by task text
manus-use history --format json | jq .   # all history as JSON
manus-use history --clear                # delete all history

History is stored at ~/.manus-use/history.jsonl.

Create ~/.manus-use/config.toml (or run manus-use init):

[llm]
provider = "bedrock"          # "openai" | "anthropic" | "bedrock" | "ollama"
model = "us.anthropic.claude-sonnet-4-20250514-v1:0"

[sandbox]
enabled = true
docker_image = "python:3.12-slim"

[tools]
enabled = ["file_ops", "code_execute", "web_search"]

[agent]
# "none" | "sliding_window" | "agentic" (model-managed, recommended for long tasks)
context_manager = "agentic"

See config/config.example.toml for all available options.

AWS Bedrock:

[llm]
provider = "bedrock"
model = "us.anthropic.claude-sonnet-4-20250514-v1:0"
# Uses ~/.aws/credentials or IAM role automatically

OpenAI:

[llm]
provider = "openai"
model = "gpt-4o"
api_key = "sk-..."    # or set OPENAI_API_KEY env var

Anthropic:

[llm]
provider = "anthropic"
model = "claude-3-5-sonnet-20241022"
api_key = "sk-ant-..."    # or set ANTHROPIC_API_KEY env var

Ollama (local):

[llm]
provider = "ollama"
model = "llama3.2"
base_url = "http://localhost:11434"

from manus_use import ManusAgent

agent = ManusAgent()

result = agent("Write a Python script that fetches weather data and saves it to CSV")
print(result)

from manus_use.agents import BrowserUseAgent

agent = BrowserUseAgent()
result = agent("Go to GitHub and find the top 5 trending Python repositories today")
print(result)

from manus_use.agents import DataAnalysisAgent

agent = DataAnalysisAgent()
result = agent("Load sales.csv, compute monthly revenue, and plot a bar chart")
print(result)

from manus_use.multi_agents import WorkflowAgent

workflow = WorkflowAgent()
result = workflow.handle_request("""
    1. Search the web for recent AI research papers
    2. Analyse the trends and create visualizations
    3. Generate a comprehensive report with insights
""")
print(result)

from manus_use.agents import VulnerabilityIntelligenceAgent
from manus_use.config import Config

config = Config.from_file()
agent = VulnerabilityIntelligenceAgent(config=config)

result = agent.handle_request("Analyse CVE-2025-6554 and create a comprehensive report")
print(result)

from manus_use import ManusAgent
from manus_use.config import Config

config = Config.from_file("path/to/config.toml")
agent = ManusAgent(config=config)

result = agent("Create a factorial function")
print(result)

ManusUse includes a multi-source vulnerability intelligence pipeline accessible via the CLI or Python API.

The manus-use analyze command runs an 8-step pipeline:

1. NVD + GitHub Advisory — official CVE metadata, CVSS, CWE
2. CISA KEV — known-exploited-vulnerabilities catalogue
3. AlienVault OTX — threat intelligence pulses and IoCs
4. PoC discovery — PoC Week trending digest, Trickest/CVE index, Exploit-DB, PacketStorm, GitHub
5. URL verification — every candidate URL is fetched and validated
6. Static analysis — code-level analysis of confirmed PoCs (network calls, payload patterns)
7. CWE correlation — weakness classification and remediation hints
8. Report generation — structured text, JSON, or Lark document output

# Analyse a CVE and print a structured text report
manus-use analyze CVE-2024-3094

# Get JSON output and extract the CVSS score
manus-use analyze CVE-2024-3094 --output json | jq .cvss_score

# Verify exploitability in a Docker sandbox, then write a Lark report
manus-use analyze CVE-2025-6554 --verify --output lark

Important: These tools are designed for defensive security purposes only. Use them for legitimate security research, vulnerability management, and defence.

• File operations (read, write, edit, delete)
• Code execution in Docker sandboxes
• Web search (DuckDuckGo, configurable)
• Browser automation (click, type, extract, screenshot)
• Data visualization (charts, plots, reports)
• Security tools (NVD, CISA KEV, OTX, Exploit-DB, Trickest, PoC Week)
• HTTP requests with content extraction
• Python REPL with persistent state

• AWS Bedrock (Claude, Titan, …)
• OpenAI (GPT-4o, GPT-4-turbo, …)
• Anthropic (Claude 3.5 Sonnet, Opus, …)
• Ollama (Llama, Mistral, … running locally)

git clone https://github.com/manus-use/manus-use.git
cd manus-use
pip install -e ".[dev,browser,search,visualization]"

# All tests
pytest tests/ -v

# With coverage
pytest tests/ --cov=manus_use --cov-report=html

# Or via hatch
hatch run test
hatch run test-cov

ruff check src/ tests/
ruff format src/ tests/

# Or via hatch
hatch run lint
hatch run format

manus-use/
├── src/manus_use/
│   ├── agents/          # Agent implementations
│   │   ├── base.py      # BaseManusAgent (all agents inherit from this)
│   │   ├── manus.py     # ManusAgent (general purpose)
│   │   ├── browser.py   # BrowserAgent (lightweight)
│   │   ├── browser_use_agent.py  # BrowserUseAgent (full JS automation)
│   │   ├── data_analysis.py      # DataAnalysisAgent
│   │   ├── mcp.py       # MCPAgent
│   │   └── vi_agent.py  # VulnerabilityIntelligenceAgent
│   ├── multi_agents/    # Multi-agent orchestration
│   │   └── workflow_agent.py
│   ├── tools/           # Individual tool implementations
│   ├── cli.py           # manus-use CLI entry point
│   ├── config.py        # Config model (TOML-backed)
│   └── __init__.py
├── tests/               # pytest test suite (175+ tests)
├── config/
│   └── config.example.toml
├── examples/            # Runnable usage examples
└── pyproject.toml

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

Explore the examples/ directory for runnable scripts:

• basic_usage.py — simple ManusAgent task
• browser_use_demo.py — browser automation
• multi_agent_flow.py — multi-agent orchestration

• 📖 Documentation
• 🐛 Issue Tracker
• 💬 Discussions

This project is licensed under the MIT License — see LICENSE for details.

• Built with Strands SDK — a powerful Python SDK for building AI agents
• Browser automation powered by browser-use — framework for AI-driven web automation
• Inspired by Anthropic's computer use demonstrations