
Deep Research
This is a comprehensive research workflow that turns Claude into a cited analyst. When you ask it to research a topic, it figures out what type of investigation you need (market sizing, competitive teardown, patent landscape, academic lit review, or seven others), fans out parallel web searches across multiple axes, cross-checks claims against multiple sources, and writes a prose-heavy Markdown report with inline citations and confidence flags. The core discipline is simple: every claim needs a URL or it doesn't ship. It defaults to a five-step process but adapts depth based on your phrasing. Built by samber and mcourant, requires internet access and a few CLI tools (curl, pandoc, md-to-pdf). If you've ever needed to understand a space quickly without guessing or trusting a single source, this is the skill.
- skills.sh installs
- 5.6K
- GitHub stars
- 242
- last updated
- Aug 21, 2026
Install
$ npx -y skills add samber/cc-skills --skill deep-research --agent claude-codeInstalls into .claude/skills of the current project.
Files
Persona: You are a senior research analyst. You are skeptical of single sources, obsessed with citations, and always flag uncertainty rather than papering over it.
Thinking mode: Reason as thoroughly as possible for Step 5 synthesis (standard and deep modes). Reconciling conflicting multi-source data and ranking recommendations requires deep reasoning — shallow inference produces wrong conclusions. On Claude Code, use ultrathink to trigger extended thinking explicitly.
Orchestration mode: Fan out 3–20 parallel sub-agents for research evidence gathering (Steps 2–4) — each agent owns one independent axis. On Claude Code, use ultracode to opt into multi-agent orchestration explicitly.
Modes:
| Mode | When | Execution |
|---|---|---|
| Interview | Step 1 — scope | Sequential; ask questions, confirm before proceeding |
| Parallel research | Steps 2–4 — evidence gathering | Fan out 3–20 sub-agents per step; each owns one axis |
| Synthesis | Step 5 — conclusions | Sequential + ultrathink; reconcile conflicts before recommending |
| Report writing | Step 6 — final output | Single sub-agent reads all notes, writes final report |
Sponsors
Details
- Repository
- github.com/samber/cc-skills
- Categories
- Git & Pull Requests, Code Review & Quality, Data Science & ML, Productivity & Planning, Automation & Workflows, Sales & Marketing, Finance & Trading
- First seen
- Jul 14, 2026
More from samber/cc-skills
All 21- Humaniseur Fr5.6K
- Technical Article Writer5.5K
- Copywriting Cta5.5K
- Conventional Git5.4K
- Site Launch Checklist5.4K
- Copywriting Prose Creator5.3K
- Influence And Negotiation5.3K
- Chrome Extension5.3K
- Press Release Writer5.3K
- Substack Ghostwriting5.2K
- Promql Cli5.2K
- Snyk Agent Scan Compliance5.1K
- Crxjs5.1K
- Skill Progressive Disclosure Design5.1K
- Training Report5K
- Linkedin Ghostwriter1.8K
- Copywriting Hooks6.4K
- Frontend Design Deslop5.8K
- Linkedin Ghostwriting5.8K
- Copywriting Tone Of Voice Creator5.7K