Latest Agent Skills

Routes product-discovery work into the right stage — opportunity brief, research, PRD, tickets, or outcome review — and hands off to the matching sub-skill.

Automates browser interactions for web testing, form filling, screenshots, and data extraction.

Acceptance-test newly added features using browser automation, happy-path flows, screenshots, UX notes, and a written report.

Refactors a class toward the Single Responsibility Principle by identifying responsibilities, grouping cohesive behavior, and extracting focused classes.

Refactors duplicated code by extracting helper methods, shared abstractions, or reusable structures, with automated testing and commits after each extraction.

Runs a UI redesign experiment by creating three git worktree branches for modern minimal, bold creative, and classic refined design directions, then implements and compares each approach.

Runs a ten-round adversarial gauntlet to pressure-test absolute claims, refining them with explicit boundaries and an Oracle synthesis.

Reviews development plans for gaps, hidden assumptions, critical misses, and XP violations, then presents prioritized findings for confirmation.

Discovers and installs agent skills from the open ecosystem using the Skills CLI, helping users find relevant skills for their tasks.

Refines an existing skill in place with minimal diffs, using evidence from usage and session mining to improve triggers, workflow, metadata, and resources.

Audits over-engineered codebases by factoring layers into live obligations, quotienting redundant distinctions, ablating unearned surface, and normalizing survivors while preserving required behavior. Produces an evidence-backed Reduction Certificate, cuts, migration phases, proof signals, and rollback.

Mines a codebase for evidence-backed improvement opportunities, forcing two escalation gates to surface breakthrough ideas, and outputs a ranked portfolio with a plan seed without implementing.

Forces a second, harder escalation pass using the original glazer prompt words verbatim, pushing for a sharper replacement thesis rather than more polish.

Remove AI writing patterns from prose by cutting filler phrases, breaking formulaic structures, and enforcing active voice and specificity.

Interviews the user in depth about a plan, probing technical implementation, UI/UX, risks, tradeoffs, and missing decisions before writing or refining a spec.

Generates a concise bullet-point summary of branch changes for a pull request description, grouping related commits by theme and verifying details against actual files.

Writes user-focused acceptance criteria for Jira stories using Given/When/Then scenarios, and updates the ticket via the Atlassian MCP server.

Analyzes a local codebase to infer its actual architecture from code-first evidence, producing a dominant architecture label, coexisting patterns, drift, and repo-fit advice for downstream agents.

Forces one deeper, more creative pass on an answer or plan while preserving the original escalation prompt verbatim.

Extracts a language-agnostic ghost package (spec, tests, install and verify docs) from an existing repository, preserving behavior via tests.yaml and evidence bundles.

Compiles up to five durable insights from the current conversation for possible addition to CLAUDE.md, focusing on patterns, preferences, and project knowledge.

Clarifies ambiguous or conflicting requests by researching first, then asking only judgment calls. Use when prompts say "grill me", ask hard questions, request relentless interrogation, pressure-test assumptions, clarify scope/requirements, define success criteria, or request system-design/optimization decisions.

Delivers measurement-driven performance optimization for latency, throughput, memory, and tail behavior, with correctness preservation and regression guards.

Generates a comprehensive ONBOARDING.md for a codebase, covering architecture, key files, setup, and common gotchas.

Rebuilds working context after a session clear by reading git history, plans, and beads issue state, then produces a visible briefing.

Reviews and repairs risky or unclear code with minimal validated changes, running a fixed-point loop until the diff is clean across two consecutive review cycles.

Apply strict-mode Bash scripting conventions: minimal validation, concise logic, executable scripts, and no unnecessary comments or echoes.

Performs structured code reviews of git branches, tracking findings in a living document and optionally running deep architectural analysis with subagents.

Stages and commits changes with quality checks, status review, and a clear message, then shows recent history.

Guides minimal-diff code changes by defining contracts and invariants, choosing a stable seam, and proving the result with an executed check.

Reviews a codebase for structural smells and refactors them with provable safety, using micro commits and green-bar discipline.

Decompose large user stories into thin, independently deliverable vertical slices using the SPIDR method (Spike, Path, Interface, Data, Rules).

Converts existing PRDs into the prd.json format used by the Ralph autonomous agent system, including splitting large features into right-sized stories and creating a progress file.

Refactors production code with test coverage using a TDD workflow that applies approved refactorings one at a time, gates each on passing tests, and commits or reverts accordingly.

Saves the latest learning as a numbered markdown file under ~/.claude/learnings and updates the learnings index with title and tags.

Analyzes Claude Code session logs to report token usage, tool and skill rankings, and session search by prompt text.

Enforces strict outside-in red-green-refactor TDD with one-failure-per-turn discipline, predicting failures before every run, hardcoding minimum changes, and using triangulation and property-based testing. Triggers on explicit TDD requests and tasks phrased as 'implement', 'add', 'write a function', 'build a feature'.

Generates a structured Product Requirements Document (PRD) for a new feature, including clarifying questions, user stories, functional requirements, and success metrics.

Mantém a coerência visual entre peças geradas por IA usando modelo fixo, prompt base, seed e referência de estilo, com teste de coerência e biblioteca de prompts.

Transforma copy aprovada em conceito visual: ideia central, mecanismo visual, hierarquia, sistema de peças e especificação de ativos com prompts prontos.

Monta carrosséis completos para Instagram, LinkedIn e Facebook, estruturando slides, definindo hierarquia e construindo no Figma como componente com variantes, com revisão e texto alternativo.

Audita criativos antes da aprovação, verificando contraste, legibilidade, área segura, hierarquia, aderência ao brand kit e conformidade, com veredito de aprovação, ajuste ou reprovação.

Cria e mantém o contexto de criativos da marca, incluindo Figma, brand kit, formatos, geradores e restrições, para uso por outras skills do plugin.

Composes final brand assets in Figma from generated images, applying text, brand tokens, layout, and grid. Optional layer: use when Figma is connected and an editable file is needed. Without Figma, delivers a composition specification.

Exports approved Figma assets and assembles the delivery package with named files, captions, alt text, and publishing instructions.

Organiza, compara e seleciona variações de imagens geradas, publicando uma grade de comparação como Artifact para aprovação do time, com critérios de avaliação e registro de decisões.

Avalia a confiabilidade de uma plataforma inteira, mapeando pontos únicos de falha, calculando error budgets a partir de SLOs, revisando prontidão para produção e analisando resiliência e capacidade.

Benchmark, load test, capacity plan, and cache with k6, JMeter, Locust, and pgbench. Use when the user says "slow", "performance", "load test", "stress test", "how many users can it handle", "capacity", "cache", "benchmark", "k6".

Aumenta a resolução e nitidez de imagens via API Freepik/Magnific quando uma chave está configurada, ou orienta o processo manual e alternativas locais quando não há chave, sempre exigindo confirmação antes de gastar crédito.

Diagnoses and operates Kubernetes clusters: pods, deployments, ingress, HPA, PVC, StorageClass, NetworkPolicy, Secrets, Helm, ArgoCD, cert-manager. Reads before changing; mutations require confirmation and diff review.

Automates repetitive infrastructure and deployment tasks with Terraform, OpenTofu, Ansible, GitHub Actions, and n8n, enforcing safe practices like plan review and confirmation before applying changes.

Design precise, production-grade enterprise UI — dashboards, admin panels, SaaS interfaces, and data-heavy applications. Use this skill for professional software interfaces that need Jony Ive-level craft: clean, minimal, functional, with intentional personality. Do NOT use for creative landing pages, marketing sites.

Estrutura prompts para geradores de imagem e vídeo, otimizando custo e consistência com a marca.

Fornece dimensões, proporções, áreas seguras, tipografia mínima e contagem de slides por plataforma, consultando uma tabela de referência. Use quando precisar de especificações de formato para mídias sociais.

Estima e acompanha o custo de geração de ativos pagos, orientando a triagem, a confirmação e o controle de gastos em campanhas.

Cria peças verticais 9:16 (story, reels, shorts, TikTok) respeitando as áreas seguras da interface do app, com guias, legendas acessíveis e verificação por screenshot.

Cria peças únicas de feed (quote card, card de dado, capa, anúncio estático) com hierarquia visual e boas práticas de design, integrando-se ao Figma para montagem e screenshot.

Conduz investigação de incidentes com método: triagem, hipóteses, mitigação, RCA e postmortem, priorizando evidências e confirmação humana.

Otimiza imagens Docker com multi-stage builds, reduz tamanho, acelera builds e aplica práticas de segurança, incluindo análise de vulnerabilidades com Trivy.

Canonical design token registry covering colors, typography, spacing, sizing, border radius, shadows, z-index, breakpoints, motion, and focus tokens, with a CSS variable template for project setup.