Coding & Development

Inspect repository meta-knowledge integrity, verify 100% boundary coverage, detect architectural drift, and opportunistically backfill missing facts without re-initialization.

Run /test to write a test suite for code you just built or changed, after implementing a feature, route, or fix. Targets uncommitted changes automatically, reads test preferences.json for your framework (asks and saves it if absent), and picks the right strategy per file: happy path, edge cases, error states, accessibility.

Run /sync as the last step after a change is complete, around merge, to keep durable knowledge current. Updates root and nested AGENTS.md, reconciles the scope from repo evidence, and flags specs the change made stale. Surgical edits only: it adds lines, and rewrites single lines it owns. Never a whole section, never curated prose.

Run /develop to build a feature, UI or backend, from an approved design, a page, component, API, service, or data slice. If something load bearing is undecided and no spec records it, it stops and routes you to /architect; otherwise it reads the spec plus AGENTS.md, builds, and advances the scope.

Use when writing, adding, changing or reviewing tests, when fixing a bug that needs a regression test, or when auditing or pruning an existing test suite. Gates every new or changed test on the behavior it protects, rejects duplicate or implementation-coupled proof, removes test-only production seams, and guides evidence-based pruning of low-value tests without chasing deletion counts.

Run /architect when choosing between approaches, designing a feature or page, picking a tech stack, or when /develop says a decision is owed, anytime a load bearing technical decision is unmade. Asks deep questions, recommends an answer, and writes a build spec to docs/specs/. Owns all spec files.

Behavioral guardrails for coding agents, built on how LLMs actually fail at coding - silent wrong assumptions, overcomplication, drive-by edits to code they don't understand, and "done" without verification. Use this at the start of any non-trivial coding task - implementing a feature, fixing a bug, refactoring, reviewing a diff, or handling a vague request like "make it faster" or "clean this up" - even when the user doesn't mention guidelines. Skip only for obvious one-line fixes.

Explica código, erros e programação a principiantes totais, em linguagem simples, com analogias do dia a dia e sem jargão. Usa sempre que o utilizador disser que não percebe de programação ("sou noob", "não sei programar", "explica-me como se tivesse 5 anos", "explain like I'm five"), perguntar o que faz um código, o que significa um erro, o que é um termo técnico (API, variável, git, commit, build, servidor...), quiser aprender a programar, ou quando for claro pela conversa que é principiante —

Médico de websites — diagnostica, explica, dá opinião e corrige TUDO o que afeta o bom funcionamento de um site: erros e bugs, links partidos, formulários, alinhamentos e layout (telemóvel/desktop), imagens, velocidade, segurança, privacidade e RGPD/cookies, acessibilidade, SEO, conteúdo, compatibilidade entre browsers, domínio/SSL/email. Faz perguntas, mostra um relatório de saúde por prioridades, diz o que faria primeiro e só corrige com autorização. Usa sempre que o utilizador disser "o meu s

Convert text into spoken audio AND transcribe audio/video back into text and subtitles using the Speakee API. This skill does BOTH text-to-speech (TTS: speak, read aloud, voice-over, audio from text) AND speech-to-text (STT: transcribe audio, generate subtitles, SRT/WebVTT, captions, transcript from a recording). Use it whenever a user asks to speak, read aloud, pronounce, voice-over, generate audio/speech/MP3 from text; wants text-to-speech, TTS, voice synthesis, or a sound file; asks which voi

Investigates and fixes software bugs through evidence-based root-cause analysis. Reads stack traces, logs, and error messages; traces logic, runtime, concurrency, security, and performance defects across files; proposes minimal fixes with a regression review and targeted tests. Use when the user reports a bug, crash, exception, failing test, wrong output, flaky behavior, or slowdown, or asks to debug code, find bugs, or review code for defects in Python, JavaScript, TypeScript, Java, C, C++, C#,

Find and remove redundant C++ unit tests, then fill coverage gaps by analyzing gcov reports and adding targeted tests — all verified with before/after coverage comparison

Compress MP4 files to smaller sizes while preserving visually lossless quality by intelligently reducing framerate and re-encoding with optimal codec/CRF settings

Find and remove redundant Rust unit tests, then fill coverage gaps with targeted tests - all verified with before/after cargo llvm-cov comparison.

6698 sayılı KVKK (2024/2026 reformları dahil), AB GDPR (General Data Protection Regulation), 6563 sayılı ETK (İYS), 6502 sayılı TKHK (Mesafeli Sözleşmeler) ve 5651 sayılı Kanun kapsamında Privacy by Design & by Default mimarisi, veri güvenliği (RLS, şifreleme, PII maskeleme), aydınlatma ve açık rıza mekanizmaları, çerez uyumu, yurt dışı veri aktarımı ve 72 saatlik veri ihlali müdahale kurallarını içeren kapsamlı yapay zeka ajan yönergesi. AI ajanları web, mobil, SaaS ve yapay zeka uygulamaları g

Médico de aplicações — diagnostica, explica, dá opinião e corrige TUDO o que afeta o bom funcionamento de uma app (iOS, Android, Flutter, React Native, Expo ou desktop): crashes e bugs, fluxos partidos, alinhamentos e layout em todos os ecrãs, ícones e imagens, velocidade, memória e bateria, segurança, privacidade (permissões, Privacy Manifest, etiquetas de privacidade, Data safety, RGPD), acessibilidade, compras na app, e as regras da App Store e do Google Play antes de enviar para revisão. Faz

Add svgbob ASCII diagrams (```svgbob) to doc comments for Rust projects using the svgbobdoc crate

Token-economy working mode for coding agents — raw data never enters the context, only the result does. Use when running tests, reading logs, querying databases, inspecting large files or JSON, checking many similar things, doing browser checks, or when the user says "f-mode", "filter first", "save tokens", "context is filling up", "too much output". Covers filtering without hiding failures, compact summaries, verification of summaries, and when to delegate to a subagent.

Use when a codebase built quickly with AI coding tools has become hard to change or trust - competing ways of doing the same thing, duplicate or _v2/_final files, tests that pass but prove nothing, dead code, fixes that break unrelated features, code nobody can explain - or when asked to audit, clean up, untangle or de-slop a repository, or to make it safe for coding agents to keep working in.

Turns an idea or half-written requirement into one consolidated REQUIREMENTS.md plus offline mockups (UI screens, API reference, or CLI transcripts). Researches the codebase and how established products solve the problem first, then asks only the product and technical questions that change what gets built.

Use when writing, modifying, refactoring, or reviewing code. Trace a caller's operation through its API, application code, middleware, and database to cut redundant validation, round trips, and data movement.

Explains why Go module dependency versions changed between two revisions (a git ref or two directories). For every module that was raised, added, lowered or removed in the build list, it names the dependency that required the new version and the chain back to go.mod, and it lists the dependencies that forced the go directive up. Use when reviewing a Dependabot or Renovate pull request for a Go module, when go get or go mod tidy changed more modules than asked, when the go line in go.mod went up

Acts as a senior Flutter architect. Use whenever creating, scaffolding, extending, refactoring or reviewing a Flutter/Dart app or feature, or when turning a design (Figma, screenshot, mockup) into Flutter screens. Gathers requirements before writing code, then enforces feature-first layered architecture, design-system-first UI, zero magic values, localization, typed routing, layered networking, and a self-review gate that must pass before delivery.

Runs, debugs and fixes Playwright end-to-end tests through the playwright-e2e-mcp server: run results grouped by root cause, trace-based failure diagnostics, verified locators, self-healing fixes that re-run and revert if still red, a heal guard that blocks fixes which pass by checking less, per-test flaky verdicts (parallelism, test order, timing), an owned, expiring flaky-test quarantine, saved logins with 2FA codes, signed-in page inspection with network mocks, tests written by walking the flow, CI report diagnosis, accessibility and Web Vitals audits and run history with the commit that broke each test. Use when the user asks to run Playwright tests, fix a failing or broken e2e test, find out why CI failed, find flaky, unstable or slow tests, repair a selector, write or review a Playwright test, log in for tests, or check a page for accessibility, performance or visual regressions.