Research skills for Claude Code, Codex, Cursor, and meshcode — deep research, literature review, customer research, survey design, patent landscape analysis, competitor research, dossiers, and research briefs. The 8 research skills (res-*) are distilled editions containing judgment knowledge only, with no scripts or install procedures: source credibility tiers (primary > official reports > industry > press), literature screening criteria, survey sample-size formulas, patent IPC/CPC analysis methods, and a fixed output contract (## Output) for every skill. They follow the Agent Skills standard (agentkills.io) — just unzip them into your project's .meshcode/skills/ in a flat layout, and Claude Code, Codex, Cursor, and meshcode desktop will pick them up automatically from the next session onward.
- PMs and planners — quickly turn competitor analysis, customer research, and survey plans into evidence-backed documents
- Researchers and consultants — literature review and patent landscapes as reproducible procedures
- Startup founders — company and people dossiers before investor meetings, validating market hypotheses with surveys
- AI agent operators — a collection of knowledge-first skills ready to publish to stores and registries
| Skill | Role | Core judgment criteria |
|---|---|---|
res-deep-research |
Multi-source synthesis research | Credibility tiers 1–4, conflict resolution rules |
res-litreview |
Literature review | PRISMA screening, finding-level synthesis |
res-customer-research |
Customer & user research | JTBD translation, 5 bias checks |
res-survey-design |
Survey design | Sample-size formula (n≈384), non-response bias |
res-patent-landscape |
Patent landscape | IPC/CPC classification search, CR5 · 3 FTO elements |
res-competitor-research |
Competitor analysis | Complaint frequency, moat judgment, price normalization |
res-dossier |
Company & people dossiers | [F/E/I] tags, mandatory unverified section |
res-research-brief |
Research plans | Decision sentence, success criteria agreed in advance |
- Download the zip → extract into your project's
.meshcode/skills/(keep the folder flat:.meshcode/skills/res-deep-research/SKILL.md) - Start a new Claude Code · Codex · Cursor · meshcode session → skills are exposed automatically
- Full catalog at the hub: github.com/meshcode-ai/skills
Multi-source deep research with citation-grade synthesis — frames the question, gathers primary/official sources, resolves conflicts, delivers a sourced verdict. Use when the user says "deep research", "research this thoroughly", "investigate this for me", "research request"…
Structured literature review — search strategy, inclusion/exclusion screening, quality appraisal, and synthesis map of a research field. Use when the user says "literature review", "prior work review", "literature search"…
Customer and user research — interview guides, JTBD framing, evidence-based insight synthesis, and bias controls. Use when the user says "customer research", "user research", "customer interview"…
Survey and questionnaire design — sampling plan, sample-size math, question wording, order and non-response bias checks, analysis-ready instrument. Use when the user says "survey design", "questionnaire design", "sample size"…
Patent landscape and freedom-to-operate analysis — IPC/CPC classification mapping, assignee and citation clustering, white-space and risk screening. Use when the user says "patent landscape", "patent analysis", "FTO"…
Competitor research — systematic profiling of rivals' product, pricing, positioning, GTM, and moat with evidence-graded comparison tables. Use when the user says "competitor analysis", "benchmark"…
Structured dossier building — entity briefing packs on companies or people with verified facts, timeline, network map, and open questions. Use when the user says "dossier", "company profile", "person research"…
Research brief writing — turns a fuzzy topic into a scoped, decision-oriented brief with questions, success criteria, and an executable plan others can run. Use when the user says "research brief", "research plan"…
- Hub catalog: github.com/meshcode-ai/skills (llms.txt · index.json · robots.txt)
- skills-seo — 7 SEO/AEO skills
- Planned: skills-marketing · skills-copy · skills-ops · skills-biz · skills-exec (rolling releases under the same spec)
These skills are built for meshcode (free download — macOS/Windows):
- Open your project in meshcode
- In chat, ask "show available skills", then "install the research skills" — meshcode fetches from this repo automatically, no git or terminal needed
- They appear in the next session and load only when a task matches, so installing all of them stays cheap
Manual alternative: download this repo's zip and extract into your project's .meshcode/skills/. Also works in Claude Code (~/.claude/skills/), Codex, and Cursor.