manus

What This Skill Does

The Manus skill provides a direct interface for OpenClaw users to leverage the Manus autonomous AI agent platform. By utilizing the provided scripts/manus.sh utility, users can offload complex, multi-step tasks—such as comprehensive web research, data synthesis, and complex reasoning—to the Manus infrastructure. The skill handles the complete task lifecycle: submission, status monitoring, result retrieval, and file management. It acts as a bridge, allowing your local environment to orchestrate sophisticated agentic workflows while maintaining strict control over API key management and task triggering.

Installation

To integrate this skill, ensure you have curl and jq installed on your system. Run the following command in your terminal: clawhub install openclaw/skills/skills/mvanhorn/manus After installation, obtain an API key from manus.im and set it as an environment variable in your shell configuration: export MANUS_API_KEY='your_key_here'

Use Cases

• Deep Market Research: Ask Manus to compile competitive analysis reports, industry trends, and SWOT analyses from diverse web sources.
• Complex Content Creation: Utilize the 'max' profile for long-form, thoroughly researched technical writing or whitepapers that require cross-referencing multiple URLs.
• Automated Data Gathering: Use the agent to visit multiple pages, extract tables, and summarize findings into downloadable reports, saving you hours of manual browsing.

Example Prompts

1. "Manus, please conduct a deep research report on the current state of decentralized storage protocols and summarize the top three projects by market cap."
2. "Use the manus-1.6-max profile to create a detailed comparison table of Python vs. Go for backend microservices, including pros, cons, and performance benchmarks."
3. "Check the status of my previous task with ID 12345 and download the resulting files to the ./research-data directory."

Tips & Limitations

• Profiles: Always choose your profile wisely. manus-1.6 is standard, lite is best for quick tasks, and max is intended for deep-dive research.
• Timeout Management: Default waiting time is 600 seconds. If your research is intensive, specify a higher custom timeout in the wait command to avoid premature termination.
• Privacy: Remember that the task prompt and data are processed by the Manus API. Ensure you do not include sensitive credentials or private corporate data in your prompts.
• Manual Execution: This skill is designed to be triggered manually. It is not intended for autonomous loops, ensuring you stay in full control of your API usage and costs.