Deep Research
Run structured discovery—local repos, competitors, forums, and technical options—before committing to a new product or major feature.
Overview
Deep Research is an agent skill most often used in Idea (also Validate, Build) that produces depth-tiered research briefs before you commit to building something new.
Install
npx skills add https://github.com/jezweb/claude-skills --skill deep-researchWhat is this skill?
- Three depth modes: focused (~30 min), wide (1–2 hours), and deep (3–6 hours) with different ambition scopes
- Combines local project reuse scans with web research, competitor analysis, ecosystem/plugin signals, and technical optio
- Focused mode targets one decision with a ~1-page recommendation; wide/deep modes produce spec-ready or major-build brief
- Explicit triggers for competitive analysis, discovery, and research before coding instead of jumping from model priors
- Aims to make the subsequent build phase materially more productive via a consolidated research brief
- 3 depth levels: focused (~30 min), wide (1–2 hours), deep (3–6 hours)
- Focused depth produces a ~1-page recommendation for one decision
Adoption & trust: 755 installs on skills.sh; 841 GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You are about to build from intuition or model defaults without knowing what you can reuse, who already solved it, or which technical path fits your constraints.
Who is it for?
Solo builders starting a new SaaS feature, agent tool, or CLI who need competitor context, local reuse, and technical comparisons in one pass.
Skip if: Teams that already have an approved spec and task list—skip straight to planning or implementation skills unless a specific open question remains.
When should I use this skill?
User says research this, deep research, discovery, explore the space, what should I build, competitive analysis, before I start building, or research before coding.
What do I get? / Deliverables
You get a structured research brief matched to focused, wide, or deep scope so you can validate scope, write a spec, or start implementation with clearer options and fewer blind alleys.
- Depth-appropriate research brief (1-page recommendation through comprehensive major-build brief)
- Competitor and ecosystem notes with technical option comparisons
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Deep research is the canonical pre-build shelf in Idea because it answers “what exists and what should we choose?” before validation and implementation. The skill’s outputs are research briefs and landscape maps, which map directly to the research subphase rather than prototyping or shipping work.
Where it fits
Map direct and indirect competitors plus forum sentiment before you niche a new agent marketplace.
Run a focused depth pass to compare two editor frameworks for your docs product.
Use wide depth to turn landscape findings into MVP boundaries and non-goals for a landing experiment.
Use deep depth to evaluate plugin ecosystems and auth patterns before wiring payments and webhooks.
Identify reusable modules in local repos so the prototype reuses code instead of greenfield spikes.
How it compares
Use instead of ad-hoc chat “what should I use?” threads when you need sourced landscape notes and a brief artifact, not a one-line library pick.
Common Questions / FAQ
Who is deep-research for?
It is for solo and indie builders using Claude Code who want disciplined discovery before they invest weeks in the wrong stack, niche, or feature shape.
When should I use deep-research?
Use it in Idea (research, competitors) before you commit; in Validate (scope, prototype) when landscape work must inform MVP cuts; and in Build (integrations, pm) when a major technical or ecosystem choice still lacks evidence—especially when you say “research this,” “deep resear
Is deep-research safe to install?
Treat it as procedural guidance that may drive web and local repo exploration; review the Security Audits panel on this Prism page and constrain agent permissions for network and filesystem access during runs.
SKILL.md
READMESKILL.md - Deep Research
# Deep Research Comprehensive research and discovery before building something new. Instead of jumping straight into code from training data, this skill goes wide and deep — local exploration, web research, competitor analysis, ecosystem signals, future-casting — and produces a research brief that makes the actual build 10x more productive. ## Depth Levels The difference is **scope of ambition**, not just time. | Depth | Purpose | Scope | |-------|---------|-------| | **focused** | Answer a specific question | One decision: "CodeMirror vs ProseMirror?" — targeted search, local scan, 1-2 comparisons. Produces a 1-page recommendation. | | **wide** | Understand the space | Landscape for a new product or feature. Competitors, ecosystem, user needs, architecture options. Enough to write a spec. | | **deep** | Plan a major build | Leave no stone unturned. Everything in wide PLUS library/component research, plugin ecosystems, GitHub issues mining, community sentiment, future-casting, technical deep-dives on every decision. Enough to drive weeks of coding. | Default: **wide** ## Workflow ### 1. Understand the Intent Ask the user: - **What** are you building? (one sentence) - **Why?** What problem does it solve? Who's it for? - **Constraints?** Stack preferences, budget, timeline, must-haves? - **Existing work?** Any projects to build on? Repos to look at? If the user gives a brief prompt ("obsidian replacement on cloudflare"), that's enough — fill in the gaps through research. ### 2. Local Exploration Scan the user's machine for relevant prior work: ```bash # Find related projects by name/keyword ls ~/Documents/ | grep -i "KEYWORD" # Read CLAUDE.md of related projects for architecture context find ~/Documents -maxdepth 2 -name "CLAUDE.md" -exec grep -l "KEYWORD" {} \; # Check for reusable patterns, schemas, components find ~/Documents -maxdepth 3 -name "schema.ts" -o -name "ARCHITECTURE.md" | head -20 ``` For each related project found: - Read CLAUDE.md (stack, architecture, gotchas) - Check for reusable code (schemas, components, utilities, configs) - Note what worked well and what didn't (from git history, TODO comments) Also check: - Basalt Cortex (`~/Documents/basalt-cortex/`) for related clients, contacts, knowledge facts - `grep -rl "KEYWORD" ~/Documents/basalt-cortex/ --include="*.md"` ### 3. Web Research Search broadly to understand the space: - **Product category**: "markdown note app", "knowledge management tool for teams" - **Competitors**: find top 5-10 by searching "best X", "X alternatives", "X vs Y" - **Open source**: search GitHub for open-source alternatives, check star counts - **Architecture**: "how to build X", "X tech stack", "building X with [framework]" - **Technology docs**: check llms.txt, official docs for key technologies - **Platform examples**: "built with Cloudflare Workers", "D1 full-text search example" - **Tutorials and case studies**: "building a Y from scratch", "lessons learned building Z" ### 4. Ecosystem and Community Research (wide + deep) Go beyond the core product — the ecosystem reveals what users actually need: **Plugins and add-ons**: - What plugins exist for major competitors? The most popular ones reveal what the core product lacks. - e.g. Obsidian has 1800+ plugins — the top 20 tell you what Obsidian doesn't do well natively. - Search: "top [product] plugins", "[product] plugin directory" **GitHub