redpanda-data/ui-harness - 第 2 页

brainstorming

Use when exploring design options, starting new features, or needing to think before coding. Two modes: design (explore approaches with trade-offs) and challenge (stress-test decisions). No implementation until design is approved.

原文语言：英语

diagnose

Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce -> minimise -> hypothesise -> instrument -> fix -> regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or…

原文语言：英语

extend-harness

Extend frontend-skills hook harness. Add new rules via skill-manifest.json, pick severity tier (block/warn/nudge/info/block-strict/diagnostic), view analytics via /frontend-skills-stats. Use when adding enforcement, tuning rule severity, or debugging hook…

原文语言：英语

frontend-skills-stats

Analytics dashboard for frontend-skills hook harness. Latency percentiles, top-violated rules, zero-fire hooks, session trends. Use when user ask for hook harness stats, invoke `/frontend-skills-stats`, or want latency profiling and manifest drift checks.

原文语言：英语

grill-me

Interview the user relentlessly about a plan or design until reaching shared understanding. Phase 2b gate of /development-lifecycle. Spawns 3 parallel reviewer hats (product/engineering/design) to stress-test the plan. Use when user wants to stress-test a…

原文语言：英语

hook-audit

Analyze hook effectiveness + session retro from collected metrics. Use when user asks to audit hooks, invokes `/hook-audit`, wants to identify silent/over-aggressive/under-enforced hooks, or asks for a retro / team analytics across recent sessions.

原文语言：英语

redpanda-frontend-kit

Meta-skill that runs the generic frontend-starter-kit plus additional React skills and Redpanda-specific registry workflow. Use when bootstrapping a new Redpanda frontend project or setting up the full Redpanda frontend stack.

原文语言：英语

resolve-pr-feedback

Resolve PR review feedback by fetching unresolved threads, triaging, fixing in parallel, and replying. Use when addressing PR review comments, resolving threads, or picking up after human review.

原文语言：英语

setup-accessibility

Enforce ARIA accessibility via PostToolUse hooks -- labels, keyboard handlers, widget attributes, Playwright AXE setup. Use when setting up a11y enforcement, WCAG 2.1 AA compliance, or accessibility testing.

原文语言：英语

setup-agent-config

Token-efficient AI agent hooks -- env vars, test flag optimization, output truncation, NODE_OPTIONS. Use when optimizing Claude Code for fewer tokens or reducing context waste.

原文语言：英语

setup-atlassian-workflow

Opt-in Atlassian/Jira integration via acli -- create work items, transition status, comment, link PRs. Mirrors gh-based workflow skills for Jira users. Use when working with Jira, Atlassian, or acli.

原文语言：英语

setup-biome

Install Biome linter with Ultracite preset, create biome.jsonc config, and configure Claude Code Stop hook for auto-fix before finishing. Use when setting up linting, formatting, Biome, Ultracite, or code quality enforcement.

原文语言：英语

setup-ci-pipeline

Configure GitHub Actions CI for React/TypeScript frontend projects. Coverage gates, visual regression, caching, Blacksmith workers, bundle budgets. Use when setting up CI, optimizing pipelines, or adding quality gates to PRs.

原文语言：英语

setup-conventional-commits

Enforce conventional commit format via a PreToolUse hook on Bash that intercepts git commit commands. Replaces commitlint + husky entirely. Use when setting up conventional commits, commit message validation, or enforcing commit format standards.

原文语言：英语

setup-e2e-testing

Set up Playwright + Testcontainers + axe-core for e2e and accessibility testing. Includes patterns for forms, tables, workflows. Use when setting up e2e tests or writing Playwright tests.

原文语言：英语

setup-env-validation

Enforce type-safe environment variable access using t3-env with zod validation. Bans raw process.env.X outside dedicated env files. Use when setting up env validation, t3-env, type-safe environment variables, or banning raw process.env access.

原文语言：英语

setup-quality-gate

Add quality:gate package.json script for fast local/CI quality checks (biome + tsgo + related tests), GitHub Actions workflow, and Stop hook for type checking. Use when setting up quality gates, CI pipelines, or pre-push validation.

原文语言：英语

setup-react-compiler

Install React Compiler with rsbuild and enforce compiler-friendly patterns via PostToolUse hooks. Flags manual memoization, derived state, useRef cache. Use when setting up React Compiler or post-compiler patterns.

原文语言：英语

setup-react-doctor

React health scoring via react-doctor with Stop hook to fail on score regression. Use when setting up react-doctor or preventing UI quality regressions.

原文语言：英语

setup-react-rules

Enforce React/TS/security rules via PostToolUse hooks -- ban raw HTML, TS escape hatches, XSS vectors, barrel imports, missing passive listeners. Use when enforcing React patterns or component library standards.

原文语言：英语

setup-registry-workflow

Registry hooks + component taxonomy + consumer drift analysis. Use when maintain shadcn component registry, design system, or analyze drift between consumer repos + registry.

原文语言：英语

setup-routines

Configure Claude Code routines for automated PR review, codebase health, issue triage, and docs drift detection. Use when setting up recurring automation, GitHub-triggered workflows, or API-triggered tasks that run on Anthropic cloud infrastructure.

原文语言：英语

setup-toolchain

Enforce bun + tsgo as toolchain via PreToolUse hooks. Blocks npm, npx, tsc, global installs. Use when setting up toolchain enforcement or banning npm.

原文语言：英语

setup-ux-copy

Enforce UX text style via PostToolUse hooks. Two scripts -- ux-copy-check.sh for `.ts`/`.tsx` UI strings (bans exclamations, "successfully", "click here", Yes/No buttons, Title Case), and prose-style-check.sh for `.md`/`.mdx` prose (bans em dashes, AI-tell…

原文语言：英语

setup-zustand

Enforce zustand best practices via PostToolUse hooks -- double-parens create, useShallow selectors, persist middleware. Use when setting up zustand enforcement or preventing re-render issues.

原文语言：英语

steelman

Argue strongest case against the user's assertion with evidence. Explicit anti-sycophancy guard. Use when user says "steelman", "/steelman", "give me pushback", "second opinion", "am I wrong", or for high-stakes/irreversible calls where wrong premise costs…

原文语言：英语

tdd

Test-driven development with red-green-refactor loop. Use when writing tests, creating new features, or fixing bugs. Includes planning phase, tracer bullets, async leak detection, deep module design, and condition-based waiting.

原文语言：英语

work

Alias for /development-lifecycle. Full frontend workflow: understand -> plan -> grill -> implement (TDD) -> /go (verify -> ship). Use when doing any React/TypeScript/UI development task.

原文语言：英语

write-a-skill

Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill.

原文语言：英语

zoom-out

Go up a layer of abstraction and map the broader context. Use when unfamiliar with a code area, need to understand how something fits into the bigger picture, or want a high-level overview.

原文语言：英语