langgenius

Use when editing or auditing the Service API specs in the Dify docs repo ({en,zh,ja}/api-reference/openapi_service.json), or on any task touching API endpoint parameters, responses, error codes, status codes, or examples.

Use when preparing documentation updates for a Dify release. User provides two version references to compare (e.g., v1.13.2 and v1.13.3, or v1.13.2 and main). Covers API reference, help documentation, environment variable changes, and UI i18n glossary impact.

Use when writing, rewriting, or auditing environment variable documentation for Dify self-hosted deployment. Applies to en/self-host/deploy/configuration/environments.mdx. Covers the full process from codebase tracing to user-facing descriptions.

Check formatting compliance in changed Chinese and Japanese documentation against writing-guides/formatting-guide.md, tools/translate/formatting-zh.md, and tools/translate/formatting-ja.md. Use after finalizing a translation batch, or when the user says "check formatting (zh/ja)", "check CJK formatting", or "format audit" on translated content.

Check formatting compliance in changed English documentation against writing-guides/formatting-guide.md. Use after finalizing a draft, or when the user says "check formatting", "format audit", "format check" on English content.

Use when drafting or editing any page under `en/cli/` in dify-docs. Covers reader segments, the writing rules, content ownership, and codebase-verification rules for the Dify CLI (`difyctl`) doc set. Triggers: "draft a CLI page", "write difyctl docs", "edit a page in en/cli/".

Research a Dify feature before writing or optimizing documentation. Use when starting any doc task that requires understanding a feature's implementation, user pain points, or community feedback. Triggers: 'research this feature', 'investigate the code for', 'what do users say about', 'let's understand how X works before writing', or any documentation task where the current docs are being rewritten or significantly expanded.

Use when writing, improving, or reviewing Dify user guide documentation. Covers pages in en/{cloud,self-host}/use-dify/, en/develop-plugin/, and en/self-host/deploy/. Triggers: "write docs for [feature]", "improve this page", "review this documentation section".

Review Dify frontend code for correctness, accessibility, component design, dify-ui usage, data/query boundaries, performance, and tests. Trigger for `.tsx`, `.ts`, `.js`, UI, React, Next.js, pending-change, or focused frontend review requests.

Use when writing, refactoring, or reviewing React/TypeScript components in Dify web, especially decisions about component ownership, props/types, URL/query state, Jotai state, async state, generated API contracts, queries/mutations, overlays, effects, navigation, performance, and empty states.

Write, update, or review Dify end-to-end tests under `e2e/` that use Cucumber, Gherkin, and Playwright. Use when the task involves `.feature` files, `features/step-definitions/`, `features/support/`, `DifyWorld`, scenario tags, locator/assertion choices, or E2E testing best practices for this repository.

Summarize user input for Agent Builder E2E coverage.

Generate Vitest + React Testing Library tests for Dify frontend components, hooks, and utilities. Triggers on testing, spec files, coverage, Vitest, RTL, unit tests, integration tests, or write/review test requests.

Lightweight coding guardrails for making focused, simple, and verifiable changes in this repo. Use for all coding work.

Review backend code for quality, security, maintainability, and best practices based on established checklist rules. Use when the user requests a review, analysis, or improvement of backend files (e.g., `.py`) under the `api/` directory. Do NOT use for frontend files (e.g., `.tsx`, `.ts`, `.js`). Supports pending-change review, code snippets review, and file-focused review.

Review Mosoo changes for project-rule, architecture, type-safety, data-scope, generated-source, observability, dependency, and verification violations. Use before committing, after completing a feature, or when reviewing another agent's diff, especially for API, GraphQL, D1, runtime, auth, UI, or infrastructure changes.

Avoid unnecessary React effects in Mosoo while using React useEffect directly for real external synchronization. Use when writing or reviewing React components, refactoring effect-driven derived state or event relays, handling subscriptions and browser APIs, or validating web changes with Mosoo's focused Justfile commands.

Build or review Cloudflare Sandbox SDK integrations using repository evidence and current official documentation. Use for secure code execution, agent runtimes, container-backed tools, Wrangler bindings, Dockerfiles, preview services, or Sandbox lifecycle work, especially in an existing project whose package manager, Wrangler format, and image versions must be preserved.

Better Auth server/client setup: `auth.ts`, generated schema, DB adapters, sessions, cookies, env vars, and plugins. Use when mentioning Better Auth, betterauth, auth handlers, OAuth, email/password, or session configuration.

Scaffold Better Auth in TypeScript apps: framework handlers, DB adapters, OAuth, email/password, and auth UI. Use when adding login, sign-up, session handling, or Better Auth plugins.

Better Auth email/password setup: verification emails, password reset, policies, hashing, and credential sign-in. Use when adding or hardening email/password login, sign-up, reset, or verification flows.

Readability-first TypeScript and TSX style guardrails for strict codebases. Use when reviewing or editing TypeScript, defining shared contracts, creating lintable style rules, or deciding between type vs interface, null vs undefined, and Pick/Omit/Partial-heavy type expressions.

Act as the WTA green-impl. Use when the user is claiming a contract-sealed task, pulling its worktree, implementing the task, running checks, and submitting for review. Green-impl is the only role that edits product source code; it must stay inside the task contract's allowed files.

Act as the WTA green-lead. Use when the user is decomposing a sealed feature intent into tasks, creating a task draft, or signing a task interface contract. Green-lead owns decomposition and the implementation surface; it does not write acceptance contracts and does not implement product code.

Install the WTA Rust CLI from source. Clone the public repository, build with cargo, and place the wta binary on PATH. Use when the user wants to start using WTA but does not yet have the wta binary installed, or wants to upgrade to the latest commit on main.

Act as the WTA orchestrator. Use when the user is bootstrapping a WTA project, publishing or sealing a feature intent, registering team members or release targets, closing an intent, or running a release push. The orchestrator owns intent and release; it does not implement code or sign acceptance.

Operate WTA, the role-gated workflow controller for teams collaborating with humans and AI coding agents. Use when the user wants to set up a new WTA project, configure the SSH/VPS pair, register an identity, or follow the Blue-Green role lifecycle from intent through release.

Act as the WTA blue-harness reviewer. Use when the user is reviewing a Delivered task by pulling a reviewer worktree, running the acceptance contract's Required Checks, and recording a verdict. Blue-harness is independent of green-impl; the same person should not implement and review the same task except in solo dogfood scenarios.

Act as the WTA blue-lead. Use when the user is signing a task acceptance contract, sealing a green-lead decomposition, or merging a Green-reviewed task into product main. Blue-lead is the gate that protects product main; it does not implement code and does not run the review checks itself (blue-harness does).

Layer 3 raw HTTP: any endpoint not yet covered by a flow or resource verb.

Manage Droploft documents (drops): upload, list, get, content, share, publish, delete.

Lofts: types, switching, listing, member operations.

Droploft v2 CLI: authentication, configuration, profiles, and patterns shared across all domains.

Background task runs — list, inspect, retry. Use this to debug async failures from doc +upload, publish, etc.

Workspaces: switching, listing, members, navigation tree.

Use when writing Playwright tests, fixing flaky tests, debugging failures, implementing Page Object Model, configuring CI/CD, optimizing performance, mocking APIs, handling authentication or OAuth, testing accessibility (axe-core), file uploads/downloads, date/time mocking, WebSockets, geolocation, permissions, multi-tab/popup flows, mobile/responsive layouts, touch gestures, GraphQL, error handling, offline mode, multi-user collaboration, third-party services (payments, email verification), console error monitoring, global setup/teardown, test annotations (skip, fixme, slow), test tags (@smoke, @fast, @critical, filtering with --grep), project dependencies, security testing (XSS, CSRF, auth), performance budgets (Web Vitals, Lighthouse), iframes, component testing, canvas/WebGL, service workers/PWA, test coverage, i18n/localization, Electron apps, or browser extension testing. Covers E2E, component, API, visual, accessibility, security, Electron, and extension testing.

Fold new dev-log entries back into docs/dev-file volumes (the current-truth architecture docs), then bump each volume's 最后核对 stamp. Use when the user asks to distill/reconcile dev-file, run the weekly docs pass, or close out a workstream. Processes only the delta since each volume's stamp — cheap when run regularly.

Build or review accessible React UI. Use when implementing UI components, reviewing UI code, checking accessibility, choosing Base UI React primitives, or validating forms, overlays, keyboard behavior, focus management, accessible names, and Testing Library queries.

Use when a coding agent needs to work with Mosoo setup, local or cloud runtime state, Mosoo CLI operations, or app integration with a published Mosoo Agent.