curdx-flow
AGENTS.md4 starsChanged 6 months agoMITView on GitHub
This repository ships @curdx/flow, a Node/TypeScript CLI and Claude Code plugin bundle. src/ contains: core/ — the differentiated core (capabilities, contracts, evidence, verdict); hooks/ — hook source and shared runtime libraries (src/hooks/lib); flows/ — the optional npm bootstrap (companion picker + ~/.claude/CLAUDE.md sync) plus analyze; registry/ — the legacy companion-install layer (native claude plugin commands are primary); and i18n/, runner/, ui/ CLI plumbing. The core product lives in plugins/curdx-flow/, including .claude-plugin/plugin.json, skills/, agents/, hooks/, schemas/, templates, and references/. Generated hook…
In this repository84 other files it gives its agents.
- bmad-advanced-elicitationSkill
- bmad-agent-analystSkill
- bmad-agent-architectSkill
- bmad-agent-devSkill
- bmad-agent-pmSkill
- bmad-agent-tech-writerSkill
- bmad-agent-ux-designerSkill
- bmad-brainstormingSkill
- bmad-check-implementation-readinessSkill
- bmad-checkpoint-previewSkill
- bmad-code-reviewSkill
- bmad-correct-courseSkill
- bmad-create-architectureSkill
- bmad-create-epics-and-storiesSkill
- bmad-create-prdSkill
- bmad-create-storySkill
- bmad-create-ux-designSkill
- bmad-customizeSkill
- bmad-dev-storySkill
- bmad-distillatorSkill
- bmad-document-projectSkill
- bmad-domain-researchSkill
- bmad-editorial-review-proseSkill
- bmad-editorial-review-structureSkill
- bmad-edit-prdSkill
- bmad-generate-project-contextSkill
- bmad-helpSkill
- bmad-index-docsSkill
- bmad-market-researchSkill
- bmad-party-modeSkill
- bmad-prfaqSkill
- bmad-product-briefSkill
- bmad-qa-generate-e2e-testsSkill
- bmad-quick-devSkill
- bmad-retrospectiveSkill
- bmad-review-adversarial-generalSkill
- bmad-review-edge-case-hunterSkill
- bmad-shard-docSkill
- bmad-sprint-planningSkill
- bmad-sprint-statusSkill
- bmad-technical-researchSkill
- bmad-validate-prdSkill
- bmad-advanced-elicitationSkill
- bmad-agent-analystSkill
- bmad-agent-architectSkill
- bmad-agent-devSkill
- bmad-agent-pmSkill
- bmad-agent-tech-writerSkill
- bmad-agent-ux-designerSkill
- bmad-brainstormingSkill
- bmad-check-implementation-readinessSkill
- bmad-checkpoint-previewSkill
- bmad-code-reviewSkill
- bmad-correct-courseSkill
- bmad-create-architectureSkill
- bmad-create-epics-and-storiesSkill
- bmad-create-prdSkill
- bmad-create-storySkill
- bmad-create-ux-designSkill
- bmad-customizeSkill
- The first 60 shown.
The file
Files
Repository Guidelines
Project Structure & Module Organization
This repository ships @curdx/flow, a Node/TypeScript CLI and Claude Code plugin bundle. src/ contains: core/ — the differentiated core (capabilities, contracts, evidence, verdict); hooks/ — hook source and shared runtime libraries (src/hooks/lib); flows/ — the optional npm bootstrap (companion picker + ~/.claude/CLAUDE.md sync) plus analyze; registry/ — the legacy companion-install layer (native claude plugin commands are primary); and i18n/, runner/, ui/ CLI plumbing. The core product lives in plugins/curdx-flow/, including .claude-plugin/plugin.json, skills/, agents/, hooks/, schemas/, templates, and references/. Generated hook bundles are committed under plugins/curdx-flow/hooks/scripts/; update them with npm run build:hooks. Automation scripts live in scripts/, and tests live under tests/**/*.test.ts.
Build, Test, and Development Commands
npm ci: install dependencies frompackage-lock.json.npm run dev: runtsupin watch mode for CLI development.npm run build: build the CLI bundle todist/index.mjs.npm run build:hooks: bundle hook sources into plugin scripts.npm run typecheck: run strict TypeScript checks.npm run test:hooks,npm run test:analyze,npm run test:runner: run targeted Vitest suites.npm run test:claudecc: run the Claude Code plugin smoke test.npm run verify: run the full local release-quality gate.
Coding Style & Naming Conventions
Use TypeScript ESM targeting Node 20+. Match the existing style: two-space indentation, single quotes, semicolons, kebab-case filenames for hooks and scripts, and .ts extensions on local imports. Prefer explicit return types for exported helpers and public workflow functions. Keep plugin command names, skill folders, marketplace identifiers, and hook script paths stable unless the manifest, registry, and generated bundles are updated together.
Testing Guidelines
Vitest is the test runner. Name tests by area, for example tests/hooks/stop-watcher.test.ts or tests/analyze/parser.test.ts. Changes to src/hooks/ must run npm run build:hooks and npm run check:hooks-fresh. Plugin metadata changes should run npm run check-versions and npm run test:claudecc.
Commit & Pull Request Guidelines
Prefer Conventional Commit style seen in history, such as feat(plugin): ..., fix(installer): ..., or test(hooks): .... Pull requests should describe behavior changes, affected plugin or CLI surfaces, linked issues, and exact verification commands run. For releases, synchronize versions, run npm run verify, then push the release tags required by the plugin/npm workflow.
Agent-Specific Notes
Always respond in Simplified Chinese when working in this repository. Treat plugins/curdx-flow as the primary product surface. Companion plugins (pua, claude-mem, chrome-devtools-mcp, ui-ux-pro-max) and external MCPs (context7, sequential-thinking) are soft-detected at runtime — not hard dependencies in the plugin manifest; curdx-flow must install and degrade gracefully when they are absent. The shared capability descriptors live in src/core/capabilities/ (catalog, rules, tokens).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

