Loading listing…
Deep Research Skills
Security scan passedTwo-phase deep research: your agent drafts an outline of items and fields with you, researches each in parallel, then writes a report.
Pack · 5 skillsby Weizhena· ResearchUnclaimed · Yours? Claim it
Preview
4 images
What Deep Research Skills does
About
Deep Research Skills is Weizhena's structured research workflow for Claude Code, OpenCode and Codex, inspired by the RhinoInsight paper on controlling model behaviour in deep research. Research runs in two phases, outline then deep investigation, and you stay in the loop between them. It comes in English and Chinese. What you can do with it - /research: turn a topic into an outline of items to research and the fields to collect for each - /research-add-items and /research-add-fields: extend the outline before you commit to it - /research-deep: research every item on the web with parallel agents, one by one - /research-report: turn the collected results into a markdown report with a table of contents Good for - Academic research: paper surveys, benchmark reviews and literature analysis - Technical research: comparing technologies, evaluating frameworks and choosing tools - Market research: competitors, industry trends and product comparisons - Due diligence: company research, investment analysis and risk assessment How to install it Clone github.com/Weizhena/deep-research-skills, copy the skills (English or Chinese) into your agent's skills folder, install the web-search agent and its modules, and pip install pyyaml. OpenCode needs OPENCODE_ENABLE_EXA=1 for web search. Codex has its own skill variants and a script that updates ~/.codex/config.toml. Good to know - No sign-in: the research runs on your agent's own web search. - MIT licensed.
More from Research
Skills people install alongside Deep Research Skills.
