Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Analyzes technical papers and codebases to produce structured reports in Markdown and PDF formats. Supports two modes: lite for analysis and writing only, full for also setting up environments and running experiments. Iteration count controls how many revision passes the agent makes before finalizing.
Combining literature search, code reading, iterative drafting, and PDF export into one command removes the need to coordinate separate tools for each step.
When to use it
- Summarizing an arXiv paper into a readable technical report
- Reviewing an open-source ML project's architecture from its source code
- Generating a PDF write-up of experiment results for a research team
- Producing a multi-section report on a paper with citations and diagrams
- Documenting a local codebase with method explanations and a technology map
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: research-report --topic "VGGT" --iterations 5 --mode lite --output both
- 1Searches arXiv and project pages for the VGGT paper and related references
- 2Fetches citations and maps key technologies mentioned in the paper
- 3Drafts a report outline and writes each section over 5 iterative revision passes
- 4Adds Mermaid or ASCII diagrams to illustrate architecture
- 5Calls the md2pdf skill to export the final Markdown as a PDF
A versioned report directory containing VGGT_report_v1.md through VGGT_report_final.md and VGGT_report_final.pdf
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.