Testing Skills

> Proactive bug-hunting methodology — autonomous codebase reconnaissance to find latent bugs, hidden failure modes, and ticking time bombs BEFORE they detonate in production. Use this skill whenever the user asks to "find bugs", "hunt for issues", "audit this codebase", "what could go wrong", "find what's broken", "stress-test this project", "find landmines", "what will bite me later", or any variation of proactive defect discovery. Also trigger when the user says "scan for problems", "pre-flight check", "what am I missing across the codebase", "find the weak spots", or asks for a proactive quality sweep before a release, merge, or deployment. This is NOT for debugging known errors (use systematic-debugging) or reviewing specific code you're handed (use hostile-review). This is for when there are no known errors yet and you need to go FIND them.

Generate thorough, interactive HTML testing checklists that help developers and testers systematically verify software — a real test plan organized into end-to-end flows, with a prominent progress bar, pass/fail/blocked step states, filter and search navigation, and syntax-highlighted code and command snippets. Always ends with a Submit button (calls `submitToClaude`) so results come back to the agent for failure triage and fixes. Use whenever the user wants to test, QA, verify, or validate a change, feature, release, PR, or bug-fix batch — "help me test this", "test plan", "QA checklist", "verification checklist", "smoke-test list", "regression checklist". Bug lists in issue trackers (Monday, Linear, Jira, GitHub…) are natural input — render the open items as a two-way checklist whose rows keep ticket ids, and offer to write verdicts back to the tracker after Submit. Every embedded snippet passes a redaction step so credential-shaped values never land in the artifact.

Preserve visual continuity across video scenes with previous-final-frame starts, camera swipes, carryover frames, transition QA, and boundary proofing. Use when a scene must begin from the exact last frame of a previous scene, when adding camera-move transitions, when replacing scene endings, or when debugging stale starting frames.

Verify final video renders and make tiny safe edits without breaking approved sync, including proof frames, contact sheets, media probing, stale theme scans, exact timestamp checks, carryover-frame refreshes, and re-render-only-affected-scenes workflows. Use when a user asks for final QA, HD/render verification, a small text/audio/color change, or regression checks after approved timing.

用精简流程帮你从想法进入 spec、design、build，并按需执行 QA 与发布门禁。

Tests in real browsers. Use when building or debugging anything that runs in a browser. Use when you need to inspect the DOM, capture console errors, analyze network requests, profile performance, or verify visual output with real runtime data via Chrome DevTools MCP.

Guides systematic root-cause debugging. Use when tests fail, builds break, behavior doesn't match expectations, or you encounter any unexpected error. Use when you need a systematic approach to finding and fixing the root cause rather than guessing.

Use when building, modifying, or reviewing user-facing UI, responsive states, accessibility, visual fidelity, functional QA, monkey testing, or product-grade frontend polish.

Optimizes application performance. Use when performance requirements exist, when you suspect performance regressions, or when Core Web Vitals or load times need improvement. Use when profiling reveals bottlenecks that need fixing.

Get Codex's review of an implementation plan before the user starts building — trigger when they want Codex (or a named GPT model like gpt-5.6) to review, critique, or pressure-test a plan. Applies to any plan-shaped artifact — a plan file, plan-mode plan, a migration/rearchitecture/integration/checkout write-up, or a doc describing how they intend to do something (pasted, at a repo path, or just described). Fire on any second-opinion phrasing — "poke holes in it", "sanity-check my plan", "run this past Codex", "get Codex's take", "flag the biggest risks", "look this over before I start", "did I miss edge cases, testing, or a rollback path?". The point is catching gaps, risks, missing steps, and better alternatives ahead of implementation. Do NOT use for reviewing already-written code or diffs (that's diff-review), mapping an existing codebase's architecture (codebase-analysis), or web/research questions.

Generate convex-test tests for the app's Convex functions.

>- Use this skill whenever the user describes a Copilot Studio agent in a sentence or two and wants a build-ready blueprint, or asks how to design, scope, or structure an agent (its type, topics, tools, knowledge, or welcome experience). Trigger on requests like "design a Copilot Studio agent that…", "turn this use case into an agent", "how should I structure this agent", or a pasted one-line use case. Do NOT trigger for testing an existing agent (that is the test-planner skill) or for generic Power Platform flows unrelated to an agent.

Pre-simulate how a defined set of role-based personas will react to an internal comms, launch, or enablement artefact before it ships. Use when the user asks to "run the persona panel", "pressure-test this comms against our personas", "QA this launch email/deck before it goes out", "how will each team react to this", or wants persona- and domain-level feedback on a broad internal artefact. Requires a personas file — bring your own (see references/personas.template.md). Do NOT use for 1:1 private comms, HR/performance matters, or legal/contractual language.

Guides developers through a human-supervised PDCA framework for AI code generation. Applies analysis, TDD, validation, and retrospection. Activates for sessions requiring systematic quality control.

Analyze test coverage and identify gaps with actionable recommendations

Transform SDD tasks into test-first TDD task pairs. Reads existing tasks from /create-tasks and generates paired test tasks with RED-GREEN dependencies. Use when user says "create tdd tasks", "add tdd pairs", "convert to tdd", or wants to apply test-first ordering to SDD tasks.

Execute TDD task pairs autonomously with RED-GREEN-REFACTOR verification. Orchestrates wave-based execution with strategic parallelism, routing TDD tasks to tdd-executor agents and non-TDD tasks to standard task-executor. Use when user says "execute tdd tasks", "run tdd tasks", "start tdd execution", or wants to execute TDD-paired tasks from create-tdd-tasks.

Generate test files from acceptance criteria or existing code. Supports criteria-driven mode (from specs/tasks) and code-analysis mode (from source files). Auto-detects test framework and spawns test-writer agents for parallel generation.

Run a full RED-GREEN-REFACTOR TDD workflow for a feature. Accepts a feature description, task ID, or spec section as input. Presents a plan for confirmation then runs autonomously through all TDD phases.

Executes diagnostic investigation tasks to test debugging hypotheses. Runs tests, traces execution, checks git history, and reports evidence. (converted from agent)

Executes diagnostic investigation tasks to test debugging hypotheses. Runs tests, traces execution, checks git history, and reports evidence. (converted from agent)

Universal layered testing strategy — generate a complete test plan, layered architecture, and execution plan based on project type (Backend+APP / Backend+WEB / Backend+APP+Embedded)

Discover missing UAT test scenarios from the user's perspective, confirm requirements through guided interviews, and output user story files in story-craftsman template + Gherkin Given/When/Then format. Use when the user mentions "add test cases", "missing tests for XX scenario", "write UAT cases", "write test scenarios", "brainstorm test cases", or discovers test coverage gaps. Should also trigger even if the user just mentions a feature "hasn't been tested for XX situation".

Downloads, exports, and inspects existing Arize traces and spans to understand what an LLM app is doing or debug runtime issues. Covers exporting traces by ID, spans by ID, sessions by ID, and root-cause investigation using the ax CLI. Use when the user wants to look at existing trace data, see what their LLM app is doing, export traces, download spans, investigate errors, or analyze behavior regressions.

Use when an implemented web app, UI feature, workflow, or frontend change needs evidence-based validation in a real browser. Runs interactive discovery, browser execution, screenshot analysis, console/error checks, critical-flow testing, stateful resume, and HTML reporting. Supports CLI-agnostic adapters including agent-browser, playwright-cli, Playwright MCP, Claude/Codex Chrome MCP tools, and browser-harness.

Use this skill when writing new features, fixing bugs, or refactoring code. Enforces test-driven development with 80%+ coverage including unit, integration, and E2E tests.

Systematic E2E testing workflow with evidence capture and reporting. Use when running integration tests, validating complete user flows, capturing test evidence, or generating test reports.

한국 홈택스 종합소득세 신고 범용 워크플로. 모두채움/일반신고, 사업·근로·기타소득 조회, 단순·기준경비율 vs 간편장부 선택 점검, 엑셀 수입·경비 대조, 공제 누락 체크, 제출 직전 QA. 종합소득세, 종소세, 경비율, 모두채움, 5월 신고 요청 시 사용. 동료·일반 납세자용. 자동 제출·납부 금지.

>- Specialized software engineering prompt methodology for Claude. Use when building prompts for system design, code review, incident response, security analysis, API design, architecture decisions, RFCs, ADRs, review," "system design prompt," "SRE prompt," "security review prompt," "incident response prompt," "API design prompt," "architecture decision prompt," "RFC prompt," "runbook prompt." Provides 11 tested approaches across identity, reasoning, and output for engineering analysis. Do NOT use for general coding help, writing code, or debugging — this builds prompts that shape engineering analysis, not code itself. Do NOT use for evaluating existing prompts (use rootnode-prompt-validation if available).

Find and own a defensible market position. Use when messaging sounds like competitors, conversion is weak despite awareness, repositioning a product, or testing positioning claims. Includes Crawl-Walk-Run rollout methodology and the word change that improved enterprise deal progression.

Manage Apple Ads with `asc ads`: separate auth, org lookup, campaigns, ad groups, ads, keywords, reports, raw API requests, and safe live testing.

Triage failing macOS tests across Xcode and SwiftPM workflows. Use when asked to run macOS tests, narrow failing scopes, explain assertion or crash failures, or separate real test regressions from setup and environment problems.

Investigate and fix flaky tests using Tuist test insights and local repeated test runs. Use when a user provides a flaky test URL, test case identifier, or asks to find and stabilize flaky tests.

Create, stabilize, and run UI end-to-end tests with Xcode (XCUIApplication/xcodebuild), including environment setup, focus/input stabilization, logging/attachments, and flakiness triage. Use when adding or debugging UI automation, writing new UI tests, or making them reliable.

Use for slow endpoints, p95/p99 latency, database query issues, frontend bundle/render performance, memory leaks, caching, profiling, or performance regressions.

Use for test-first development, regression tests, flaky test debugging, coverage gaps, test strategy, CI failures, or converting bugs into minimal reproducible tests.

Adversarial design audit that stress-tests a game feature, system, pitch, roadmap item, or product idea by assuming failure and identifying the most credible reasons it would fail. Use when pressure-testing a concept before production, performing a pre-mortem, challenging a feature that sounds good on paper, exposing blind spots in design thinking, or getting a hostile-but-constructive critique with concrete failure mechanisms and de-risking moves.

Audit an existing or proposed premium pass (battle pass, season pass) for value proposition integrity, progression health, free track balance, reward pacing, engagement loop quality, and economy impact. Use when evaluating a pass before launch, diagnosing why a pass is underperforming, reviewing a competitor's pass design, or stress-testing a proposed structure before committing to production.

Audit a game feature, system concept, prototype plan, or preproduction proposal to determine whether the prototype is meant to sell the idea or reveal unknowns, and whether the prototype scope matches that intent. Use when teams are unclear about why they are prototyping, when a prototype risks becoming a demo in disguise, or when precious prototype time may be spent proving known strengths instead of testing real uncertainties.

Audit game design goals, feature objectives, OKRs, initiative aims, or any stated design intent against the SMART criteria adapted for game design. Use when a user shares a design goal, player experience target, retention goal, or feature aim and wants it evaluated or rewritten. Trigger on phrases like "audit my design goals", "is this a good design objective", "my feature goal is...", "we want players to feel...", "our aim for this feature is...", or whenever someone states what a design should achieve and it feels vague or unmeasurable. Also trigger when a design document contains aims or success criteria that need stress-testing. Even soft goals like "players should feel powerful" are valid — apply the framework and return both the audit and a rewritten version.

Devil's Advocate stress-testing for code, architecture, PRs, and decisions. Surfaces hidden flaws through structured adversarial analysis with metacognitive depth. Use for high-stakes review, stress-testing choices, or when the user wants problems found deliberately. NOT for routine code review. Triggers on "스트레스 테스트", "stress test", "devil's advocate", "반론", "이거 괜찮아", "문제 없을까", "깊은 리뷰", "critical review", "adversarial".

| 測試案例產生器 — 把 PRD 的驗收標準（Acceptance Criteria）與規則，轉成 QA 可直接執行的 測試案例：正常路徑、邊界值、異常路徑、併發/冪等、權限與安全。 當使用者說「幫我寫測試案例」、「這份 PRD 的 test case」、「QA 測試計畫」、「幫我補測試」、 「這功能怎麼測」、「寫測試」、「test cases」、「test plan」、「QA checklist」時， 一定要使用這個 skill。即使使用者只是丟一份 PRD、一段 AC、一個功能描述說 「幫我想想怎麼驗收」、「工程做完了要測什麼」，也應觸發此 skill。 這個 skill 輸入 PRD / AC，輸出結構化測試案例；不寫 PRD、不寫程式碼，只負責「怎麼驗證」。 接力關係：enterprise-prd-writer / prd-writer 產出 AC → 本 skill 把 AC 展開成測試案例 → QA 據以執行。PRD 的每條 Given/When/Then 是輸入，測試案例是輸出。

| 企業 PRD Writer — 受監管 / 金流 / 風控 / 跨職能團隊等級的產品需求文件撰寫與強化工具， 用於產出可直接進入 Refinement、Engineering Design、Development、QA 與 UAT 的 「施工藍圖」等級文件。相對於輕量版 prd-writer，此版本額外涵蓋權限矩陣、NFR、 依賴管理、合規、Analytics/Observability、Rollout/Migration/Rollback 與 UAT/Release Readiness。 預設以 interactive-html-report 規格輸出互動式 HTML（含常駐目錄側欄、捲動高亮）。 ⚠️ 版本選擇（重要）： 當使用者說「幫我寫 PRD」但未指明版本時，先問使用者要用「輕量版（prd-writer）」 還是「企業版（本 skill）」再開始。判斷提示：個人專案 / 單一功能 / 無合規需求 → 輕量版； 多團隊 / 金流 / 風控 / 合規 / 需上線維運全鏈路 → 企業版。使用者已明確指定版本時直接照做，不必再問。 當使用者提出以下需求時，應優先使用此 skill： 此 skill 的目標不是產出冗長文件，而是建立可執行、可測試、可追蹤、 可討論且邊界明確的產品規格，降低因需求模糊造成的返工與認知落差。 核心強制項目包含：

| 需求補洞助手（探路模式）— 專用於「白紙一張、還沒有 PRD」的場合：站在 PM、UIUX、 Backend、Frontend、QA 五個角色，掃描需求的缺漏、容易誤解的敘述、沒考慮到的 Edge Case， 以及開發前一定要確認的問題，把還沒想到的東西攤開。 ✅ 適用場合（符合任一才觸發）： 當使用者說「幫我找需求缺漏」、「這個需求有什麼沒想到的」、「補洞」、「盤點 Edge Case」、 「這是白紙設計，幫我掃一輪」、「find gaps」、「what am I missing」時，觸發此 skill。 ❌ 不適用（改用別的 skill）： 的 Gap Analysis 與 Definition of Ready，不要用本 skill 這個 skill 只負責「發散——把問題找出來」，不做格式檢查、不寫 PRD、不替使用者做決定。

| 輕量版 PRD 撰寫工具 — 適合個人專案、單一功能、無合規/金流/風控需求、單團隊開發， 快速產出可直接交付工程的「施工藍圖」等級文件。 當使用者說「幫我寫 PRD」、「產品需求文件」、「寫 spec」、「功能規格」、「write a PRD」、 「product requirements」、「feature spec」、「產品設計文件」、「需求規格書」時， 一定要使用這個 skill。即使使用者只是說「幫我整理這個功能的規格」、「把這個想法寫成文件」、 「我要交一份產品文件給工程團隊」、「寫個規格讓工程師可以直接開工」，也應優先觸發此 skill。 也適用於使用者要求「補 AC」、「加驗收標準」、「補 Out of Scope」、「補畫面狀態」等 針對既有 PRD 的強化需求。這個 skill 確保每份 PRD 都包含驗收標準、複雜度標注、 畫面狀態規格、Out of Scope 邊界，讓工程師能直接開發、QA 能直接寫測試。 ⚠️ 版本選擇（重要）：本 skill 是「輕量版」，另有「企業版（enterprise-prd-writer）」 涵蓋權限矩陣、NFR、依賴、合規、Rollout/Rollback、UAT 等。當使用者說「幫我寫 PRD」 但未指明版本時，先問要用「輕量版（本 skill）」還是「企業版」再開始。判斷提示： 個人專案 / 單一功能 / 無合規需求 → 輕量版；多團隊 / 金流 / 風控 / 合規 / 需上線維運全鏈路 → 企業版。 使用者已明確指定版本時直接照做，不必再問。

> TIA Portal TestSuite and Application Test operations. Use for managing test sets, application tests, style guide rules, and automated system testing workflows.

| UI 開発・検査・自動テストワークフロー。uitree で構造把握→pytest E2E テスト生成→PlayMode テストに移植。

Use when the user is debugging a bug, tracing an error, or asking why something fails. Examples: \"Why is X failing?\", \"Where does this error come from?\", \"Trace this bug\