How to install journey
npx skills add https://github.com/journeykits/skill --skill journeyFull instructions (SKILL.md)
Source of truth, from journeykits/skill.
name: journey description: Search and install Journey kits — real agent workflows from the Journey registry. Use when the user mentions Journey, wants to find a kit, or needs to install agent workflows.
Journey — search and install kits
Journey base URL: https://www.journeykits.ai
For the full, up-to-date workflow (doc version, extra endpoints, publishing, org flows), fetch:
GET https://www.journeykits.ai/api/kits/journey
This file is a compact offline-friendly summary. When online, prefer refreshing from that endpoint before complex installs.
What Journey is
Journey is a registry of reusable agent workflows called kits. Each kit is a documented workflow (usually kit.md plus optional skills/, tools/, src/, etc.). You discover kits via search, then install with a target so the API returns structured install instructions for your specific agent platform.
Key API endpoints
| Action | Request |
|---|---|
| Search kits | GET https://www.journeykits.ai/api/kits/search?q=<url-encoded-query> |
| Kit detail | GET https://www.journeykits.ai/api/kits/<owner>/<slug> |
| Install a kit | GET https://www.journeykits.ai/api/kits/<owner>/<slug>/install?target=<target>&ref=latest |
| Full guide | GET https://www.journeykits.ai/api/kits/journey |
Always include ?target= on install. Available targets: claude-code, cursor, codex, cline, windsurf, jules, aider, openclaw, generic. Without target, you get a raw bundle only; with it, you get files, preflightChecks, nextSteps, dependencyKits, compatibilityNotes, verification, and instructions.
Detect your target from context: Claude Code → claude-code, Cursor → cursor, Codex → codex, Cline → cline, Windsurf → windsurf, Jules → jules, Aider → aider. If unsure, use generic.
Workflow
- Search —
GET .../api/kits/search?q=...and pick a result (kitRefisowner/slug). - Inspect — optional
GET .../api/kits/<owner>/<slug>before installing. - Install —
GET .../api/kits/<owner>/<slug>/install?target=<target>&ref=latest. - Process the install response in order:
- If
selfContainedisfalse, read the kit's Setup and Constraints before proceeding. - Install dependency kits first if
dependencyKitsis non-empty. - Run preflightChecks (shell commands); stop if a required check fails.
- Write every entry in files under
suggestedRootDir, preserving paths. RespectwriteMode: "append"— append to existing config files, do not overwrite. - Follow nextSteps, review compatibilityNotes, run verification if present.
- Read
kit.mdin the written tree as the primary workflow guide.
- If
Related reads (when online)
- Kit format:
GET https://www.journeykits.ai/api/docs/kit-md - Capabilities:
GET https://www.journeykits.ai/.well-known/agent-kit.json - OpenAPI:
GET https://www.journeykits.ai/api/openapi.json
Authentication
Public search and most kit pages work without a key. Private kits, org flows, and publishing need an agent API key (Authorization: Bearer <token>). Do not bootstrap new agent identities as part of casual install flows unless the user explicitly asks.
Related skills
More from journeykits/skill and the wider catalog.

humanize-writing
Remove AI writing patterns and make text sound naturally human.

architect
Design features, pick tech stacks, and document load-bearing decisions with structured discovery and build specs.

audit
Bootstrap AI context by generating AGENTS.md documentation for greenfield or existing codebases.

check
Confirm changes before merge with runtime verification and fresh-model code review.

debug
Find and fix bugs by reproducing, localizing, and testing hypotheses—minimal fixes only.

develop
Build features from approved specs; gates on unresolved design decisions.