cookiy
AI-powered user research through natural language. Installs the Cookiy MCP server and orchestrates tool workflows for study creation, AI interviews, discussion guide editing, participant recruitment, and report generation.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
// README
Cookiy Skill
Turn Claude Code, Codex, Cursor, VS Code / GitHub Copilot, Windsurf, Cline, OpenClaw, Manus, and other MCP clients into AI user-research operators that can create studies, run interviews, recruit participants, and deliver report links from plain-English prompts.
This repository now has two layers:
- a protected production surface for the current public
cookiyskill and MCP install flow - an additive open library of prompts, references, templates, and examples for user research
The install and indexing surface stays stable while the library grows around it.
The public installer package now acts as a bootstrap CLI:
- on supported clients, it installs a local Cookiy skill copy
- then it configures MCP
- on unsupported clients, it falls back to MCP-only setup
What You Get In This Repo Today
This repo already includes concrete user-research materials, not just install docs.
| Material type | What is included now | What it is for |
|---|---|---|
| Installable runtime skill | cookiy skill + MCP install surface | Connect an agent to Cookiy and run the live study/interview/recruit/report workflow |
| Public installer package | packages/cookiy-mcp/ | Source for the cookiy-mcp bootstrap CLI and Homebrew build artifacts |
| Study-brief prompts | prompts/study-briefs/discovery-study.md | Turn a vague founder or product question into a stronger study brief |
| Interview-guide prompts | prompts/interview-guides/problem-exploration.md | Draft better exploratory interview sections, probes, and evidence goals |
| Synthesis prompts | prompts/synthesis/evidence-first-synthesis.md | Convert raw notes, transcripts, survey comments, or support inputs into findings |
| Stakeholder readout prompts | prompts/stakeholder-readouts/executive-summary.md | Turn findings into concise founder/PM/stakeholder summaries |
| Research method references | references/methods/research-method-selector.md, references/methods/thematic-analysis.md | Choose the right method and synthesize qualitative evidence better |
| Reusable templates | references/templates/research-plan-template.md, references/templates/insight-card-template.md | Standardize research planning and finding documentation |
| Book-ingestion guidance | references/books/README.md | Turn books into agent-usable notes without uploading copyrighted content |
| Worked examples | examples/study-briefs/saas-onboarding-friction.md, examples/readouts/mobile-checkout-friction.md | Show what good study briefs and research readouts look like |
| Repo governance docs | docs/PLATFORM_COMPATIBILITY.md, docs/CONTENT_POLICY.md, docs/THIRD_PARTY_ATTRIBUTIONS.md, docs/ROADMAP.md | Keep public platform compatibility stable and define how the library should grow |
Materials By Use Case
If you are here for a specific kind of material, start here:
- To plan a study:
prompts/study-briefs/discovery-study.mdreferences/templates/research-plan-template.mdexamples/study-briefs/saas-onboarding-friction.md
- To maintain the public installer package:
packages/cookiy-mcp/package.jsonpackages/cookiy-mcp/README.mdpackages/cookiy-mcp/scripts/generate-homebrew-formula.mjsdocs/contract-source.json
- To write or improve interview guides:
prompts/interview-guides/problem-exploration.mdreferences/methods/research-method-selector.md
- To analyze research data:
prompts/synthesis/evidence-first-synthesis.mdreferences/methods/thematic-analysis.mdreferences/templates/insight-card-template.md
- To write readouts for founders or PMs:
prompts/stakeholder-readouts/executive-summary.mdexamples/readouts/mobile-checkout-friction.md
- To build a learning library safely:
references/books/README.mddocs/CONTENT_POLICY.mddocs/THIRD_PARTY_ATTRIBUTIONS.md
30-Second Demo
# Install Cookiy for your AI client
npx cookiy-mcp --client codex -y
# Then ask your agent:
"Create a 6-participant checkout-abandonment study in English."
"Run 3 simulated interviews with price-sensitive mobile shoppers."
"Share the report link as soon as a preview is ready."
What happens:
- Cookiy creates the study and discussion guide inside your agent workflow.
- It runs or monitors interviews and recruitment through the MCP tool chain.
- It returns transcripts, status updates, and share links without leaving chat.
3 Real Examples
| Use case | Ask your agent | What Cookiy does |
|---|---|---|
| Ecommerce checkout drop-off | "Create a 6-participant study to understand why mobile shoppers abandon checkout after shipping costs appear." | Creates the study, generates the guide, and highlights sample size, interview mode, and duration for review. |
| SaaS onboarding friction | "Run 5 simulated interviews with first-time admins setting up SSO for the first time." | Queues AI interviews, polls status, and returns transcripts for early signal before spending on real participants. |
| Concept validation with real users | "Recruit bilingual parents in the US for 8 short interviews about a children's learning app, then share the report link when ready." | Previews targeting and pricing, launches recruitment after confirmation, tracks progress, and returns the report share link when available. |
Quick Install
Claude Code
Install as a plugin directly in Claude Code:
claude plugin add cookiy-ai/cookiy-skill
Or install the MCP server standalone:
npx cookiy-mcp --client claudeCode -y
Codex
npx skills add cookiy-ai/cookiy-skill
Or install the MCP server directly:
npx cookiy-mcp --client codex -y
OpenClaw
clawhub install cookiy # Skill only, not MCP
This installs the Cookiy skill only. To connect OpenClaw to Cookiy's live MCP tools, install the MCP server separately:
Or install the MCP server directly:
npx cookiy-mcp --client openclaw -y
VS Code / GitHub Copilot
npx cookiy-mcp --client vscode -y
Windsurf
npx cookiy-mcp --client windsurf -y
Cline
npx cookiy-mcp --client cline -y
Manus
npx cookiy-mcp --client manus -y
This writes a resumable headless OAuth helper bundle under ~/.mcp/<server>/
for Manus-style sandbox environments.
Cursor
Install from the Cursor Marketplace, or manually:
npx cookiy-mcp --client cursor -y
Any platform with Node.js
npx cookiy-mcp -y
The installer auto-detects installed AI clients. On supported clients, it installs a local Cookiy skill copy first and then configures MCP. Other clients fall back to MCP-only setup.
macOS with Homebrew
brew install cookiy-ai/tap/cookiy && cookiy -y
Security
This repository contains both skill content and an installer that writes local MCP configuration for supported clients. That means automated platform scanners may flag expected behaviors such as OAuth flows, client config updates, helper scripts, and local skill installation.
For a concrete description of the intended network access, local file writes, and credential model, see SECURITY.md.
Verification Matrix
Validation note: rows below reflect direct cookiy-mcp CLI dry-runs against the live installer contract.
| Client | Direct install command | Local validation status | Last verified |
|---|---|---|---|
| Claude Code | npx cookiy-mcp --client claudeCode -y | CLI dry-run verified, client detected locally | 2026-03-17 |
| Codex | npx cookiy-mcp --client codex -y | CLI dry-run verified, client detected locally | 2026-03-17 |
| Cursor | npx cookiy-mcp --client cursor -y | CLI dry-run verified, client detected locally | 2026-03-17 |
| GitHub Copilot / VS Code | npx cookiy-mcp --client vscode -y | CLI dry-run verified, client detected locally | 2026-03-17 |
| OpenClaw | npx cookiy-mcp --client openclaw -y | CLI dry-ru |
// HOW IT'S BUILT
KEY FILES