Explorar skills

Coordinate agent teams for adversarial review (paired skeptic/verifier per facet, challenge round with disposition vocabulary, consolidated findings with confidence) or parallel implementation (task sizing 5-6 per teammate, non-overlapping files). Enforces independent analysis before shared conclusions. Reference only (`disable-model-invocation: true`); loaded only when `agentTeams: true` in setti

Estruturas de julgamento e expressões idiomáticas essenciais para scripts Fish shell e automação. Aborda quando usar Fish vs bash, requisitos de compatibilidade macOS/Fedora e padrões específicos do Fish para prevenir bugs, como anti-padrões de variáveis universais, funções wrapper e guardas interativos.

Group sequential design methods for interim analyses, alpha spending, and futility stopping. Use when designing trials with interim looks or implementing spending functions.

Map-reduce orchestration. Parallel mappers process input chunks independently, then a dedicated reducer merges all outputs.

Foundational knowledge for writing current PyMC models including syntax, distributions, sampling, and ArviZ diagnostics. Use when creating or reviewing PyMC models.

Scaffolds production-ready legal page templates for a website: Privacy Policy, Terms of Service, Cookie Policy, Cookie Consent banner, Refund Policy, Contact / Imprint. Templates are clearly marked as a starting point — the user must customize and have a lawyer review before going live.

Run a structured organizational design health check — operationalizing the governance learning loop and decision ledger by collecting operational evidence, measuring gate effectiveness, detecting genome drift, and producing an evolution audit with routed recommendations saved to $HOME/.ai-first-kit/. Maintains the decision ledger as an append-only record. Use when the user says 'audit my design',

When setting up telemetry, debugging distributed systems, or standardizing application output.

Interactive wizard to configure platform connections, API keys, and dependencies for media-agent. Detects available tools, walks through setup for each platform, tests connections, and writes config files. Run this before any other media-agent skill.

Fan-out parallel specialists for concurrent analysis. Dispatch read-only reviewers, researchers, and analyzers against a target.

Implements features using strict test-driven development — red-green-refactor cycle. Writes a failing test first, then minimal code to pass, then refactors. Use when building features that need high reliability.

Emergency cleanup of all autoimprove experiment worktrees. Use this if a session was interrupted, Claude crashed mid-experiment, or you want to abort the loop and clean up all branches and worktrees.