manus_skill

Overview

This skill delegates complex, long-running tasks to the Manus autonomous agent for asynchronous execution. It is optimized for deep web research, multi-source synthesis, data analysis, and generating deliverables like reports, CSVs, and charts. Use this skill when tasks need parallel processing, extended runtime, or multi-step workflows that benefit from an autonomous agent.

How this skill works

You create a Manus task by sending a prompt and optional attachments or project context; Manus returns a task_id and a task_url for tracking. Poll the task status until it is completed or failed, then retrieve text outputs and any file attachments such as reports, CSVs, or visualizations. Agent profiles (lite, standard, max) let you trade speed and cost for depth and reasoning power.

When to use it

• Deep market, competitive, or product research spanning multiple websites and sources
• Stock or financial analysis that requires company research, charts, and valuation metrics
• Multi-step workflows that run for minutes to hours and benefit from autonomous parallel work
• Report generation that must include tables, CSV exports, and visual charts
• Comparative analysis across products, vendors, or feature sets

Best practices

• Specify scope, timeframe, target sources, and desired output format in the prompt
• Choose manus-1.6-lite for quick lookups, manus-1.6 for standard research, and manus-1.6-max for deep, complex analysis
• Include task_url when sharing results so stakeholders can monitor progress in the browser
• Poll status at reasonable intervals (5–10s for simple tasks, 15–30s for heavier research)
• Attach relevant files or URLs up front to ensure the agent can analyze all sources
• Track credit_usage and handle rate limits or API key errors in your workflow

Example use cases

• Create a comparative matrix of the top 5 4K monitors for Mac including specs, prices, and user notes
• Analyze a public company ticker: profile, recent performance, technical indicators, and a 1-year price chart
• Produce a market landscape report for a specific industry and region with growth trends and key players
• Synthesize product reviews and specs from multiple vendor sites into a CSV and a recommendation summary
• Run a multi-step data pipeline that ingests attachments, extracts insights, and outputs visual reports