「code mode」の検索結果

Sistema autônomo multi-agente para startups em Claude Code. Ativado com "Loki Mode". Orquestra 100+ agentes especializados em engenharia, QA, DevOps, segurança, dados/ML, operações comerciais, marketing, RH e sucesso do cliente. Vai de PRD para produto totalmente implantado e gerando receita com zero intervenção humana. Inclui ferramenta Task para despacho de subagentes, revisão de código paralela com 3 revisores especializados, triagem de problemas baseada em severidade, fila de tarefas distribuída com tratamento de dead letter, implantação automática em provedores de nuvem, testes A/B, loops de feedback de clientes, resposta a incidentes, circuit breakers e auto-recuperação. Gerencia limites de taxa via checkpoints de estado distribuído e retomada automática com backoff exponencial. Requer flag --dangerously-skip-permissions.

日本語の概要は準備中です。原文の説明を表示しています。

Visualize métricas de treinamento, depure modelos com histogramas, compare experimentos, visualize grafos de modelos e perfil de desempenho com TensorBoard - kit de visualização de ML do Google

日本語の概要は準備中です。原文の説明を表示しています。

Modelo da OpenAI que conecta visão e linguagem. Permite classificação de imagens com zero-shot, correspondência imagem-texto e recuperação cross-modal. Treinado em 400M pares imagem-texto. Use para busca de imagens, moderação de conteúdo ou tarefas visão-linguagem sem fine-tuning. Melhor para compreensão geral de imagens.

日本語の概要は準備中です。原文の説明を表示しています。

Fornece orientação para treinar e analisar Autoencodificadores Esparsos (SAEs) usando SAELens para decompor ativações de redes neurais em features interpretáveis. Use ao descobrir features interpretáveis, analisar superposição ou estudar representações monossemânticas em modelos de linguagem.

日本語の概要は準備中です。原文の説明を表示しています。

Treina modelos de linguagem grandes (2B-462B de parâmetros) usando NVIDIA Megatron-Core com estratégias avançadas de paralelismo. Use ao treinar modelos >1B de parâmetros, precisar de máxima eficiência de GPU (47% MFU em H100), ou necessitar de paralelismo tensor/pipeline/sequência/contexto/expert. Framework pronto para produção usado em Nemotron, LLaMA, DeepSeek.

日本語の概要は準備中です。原文の説明を表示しています。

Acelere a inferência de LLMs usando especulative decoding, múltiplas cabeças Medusa e técnicas de lookahead decoding. Use ao otimizar velocidade de inferência (aceleração de 1,5-3,6×), reduzir latência em aplicações em tempo real ou fazer deploy de modelos com recursos computacionais limitados. Cobre modelos draft, atenção em árvore, iteração de Jacobi, geração paralela de tokens e estratégias de deploy em produção.

日本語の概要は準備中です。原文の説明を表示しています。

Avalia LLMs em mais de 60 benchmarks acadêmicos (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use quando benchmarking de qualidade de modelo, comparação de modelos, relatório de resultados acadêmicos ou rastreamento de progresso de treinamento. Padrão da indústria usado por EleutherAI, HuggingFace e laboratórios principais. Suporta HuggingFace, vLLM, APIs.

日本語の概要は準備中です。原文の説明を表示しています。

Load this skill whenever the project supports light/dark mode, colour theme switching, high-contrast mode, or responds to prefers-color-scheme. Under no circumstances hard-code colours that break in alternative themes. Absolutely always test colour contrast in both light and dark themes, and respect user OS-level colour preferences via CSS media queries.

日本語の概要は準備中です。原文の説明を表示しています。

Load this skill whenever the project contains text, UI components, icons, form controls, data visualisations, or focus indicators — in short, almost every project. Under no circumstances hard-code colour values without verifying contrast ratios. Absolutely always ensure text meets 4.5:1, large text meets 3:1, and non-text UI elements meet 3:1 against adjacent colours. Test in light mode, dark mode, and forced-colors (high contrast) mode.

日本語の概要は準備中です。原文の説明を表示しています。

[Documentation] Use to author, audit, amend, or test-spec a business Feature Spec. The single spec skill — modes draft|init|update|audit|amend create/maintain the tech-free 8-section Feature Spec; draft authors a provisional spec from an idea/requirement (no code yet, Evidence: TBD); tests generates Section 8 TC-{FEATURE}-{NNN} test specifications; sync reconciles §8 TCs ↔ executing test code. Per-mode procedure lives in references/{author,tests,sync}.md.

日本語の概要は準備中です。原文の説明を表示しています。

Validate Commodore MPS-801 serial dot matrix printer behavior in EMULATORS, including SerialPrinterMPS801 classes and their MatrixPrinterEmulation and SerialPrinterPeripheralSimulation hierarchy, IEC/KERNAL-trap printer routing, device numbers, secondary addresses, PETSCII/character sets, 6x7 glyphs, double width, graphics mode, reverse field, tab and dot positioning, line/page geometry, PostScript output, and fidelity boundaries. Use when Codex reviews or edits COMMODORE MPS801 printer code, generated printer output, PostScript printer templates, C64/VIC-20/C264 serial-printer behavior, or compares implementation against .codex/docs/MPS801.pdf and historical MPS-801 sources.

日本語の概要は準備中です。原文の説明を表示しています。

Validate Commodore MPS-802 and 1526 serial dot matrix printer behavior in EMULATORS, including SerialPrinterMPS802 classes and their MatrixPrinterEmulation and SerialPrinterPeripheralSimulation hierarchy, IEC/KERNAL-trap printer routing, device numbers, secondary-address functions, PETSCII/character sets, 8x8 glyphs, enhanced width, reverse and paging modes, formatter strings, programmable characters, line/page geometry, PostScript output, and fidelity boundaries. Use when Codex reviews or edits COMMODORE MPS802 printer code, generated printer output, PostScript printer templates, C64/VIC-20/C264 serial-printer behavior, or compares implementation against .codex/docs/MPS802.pdf, the MPS-802 user guide, and 1526/MPS-802/4023 service information.

日本語の概要は準備中です。原文の説明を表示しています。

Validate and implement Sinclair ZX80 video emulation in the EMULATORS repository, including CPU-driven display generation, display-file execution, Z80 M1/refresh/HALT/INT interactions, character and inverse-video fetches, line counter and shift register behavior, horizontal and vertical sync, collapsed display files, PAL timing, flicker, and cassette/video coupling. Use when Codex reviews or edits namespace ZX81 code for the /mZX80 model, especially ULA, ULARegisters, MemoryVideoCode, Memory, PortManager, Screen, CPU interrupt timing, raster, or video tests.

日本語の概要は準備中です。原文の説明を表示しています。

Build an evidence-backed working model of an existing codebase. Use when you must orient in a large or unfamiliar repository; explain architecture or a subsystem; trace a request, runtime, control, or data flow for a concrete question or bug; analyze a branch, commit, diff, or pull request for behavioral impact and blast radius; or assess AI-generated or changed code against requirements, contracts, invariants, tests, and runtime evidence. Default to read-only investigation. Do not trigger only because a coding task touches an existing repository; use when understanding or verification is the requested outcome or a substantial prerequisite.

日本語の概要は準備中です。原文の説明を表示しています。

Use when images (photos, illustrations, screenshots) need to become input for TypeSafe's Jev (System One) model, which reads only text -- "turn this photo into a Jev state", "describe these images for Jev", "decode images for System One", or a Jev project whose decision depends on what a picture shows. Handles 1 to about 100 images; produces a neutral, complete decode and a JSON state, checked against Jev.

日本語の概要は準備中です。原文の説明を表示しています。

非対話セッション（claude-code の remote-control / CI など TTY が無く SSH 鍵パスフレーズを入力できない状況）で git push を通すスキル。remote が SSH URL で `ssh -o BatchMode=yes` による非対話 SSH 認証が実際に通る（agent の鍵・パスフレーズ無しのディスク鍵・macOS キーチェーン鍵のいずれでも可）なら素の SSH push を実行し、非対話 SSH 認証が通らず `git push` が `Permission denied (publickey)` で失敗する場合は、push を HTTPS + gh トークン（credential.helper='!gh auth git-credential'）経由へ自動フォールバックして標準入力なしで実行する。「push して」「リモートに反映して」と頼まれたが SSH 認証が使えない、push が publickey で弾かれた、remote-control から push したい、といった場面で使う。`/gh-push` で明示起動も可。GitHub(gh) 前提。

非対話セッション（claude-code の remote-control / CI など TTY が無く SSH 鍵パスフレーズを入力できない状況）で git fetch / pull を通すスキル。remote が SSH URL で `ssh -o BatchMode=yes` による非対話 SSH 認証が実際に通る（agent の鍵・パスフレーズ無しのディスク鍵・macOS キーチェーン鍵のいずれでも可）なら素の SSH fetch を実行し、非対話 SSH 認証が通らず `git fetch`/`git pull` が `Permission denied (publickey)` で失敗する場合は、取り込みを HTTPS + gh トークン（credential.helper='!gh auth git-credential'）経由へ自動フォールバックして標準入力なしで実行する。「fetch して」「pull して」「リモートの変更を取り込んで」「最新を取ってきて」と頼まれたが SSH 認証が使えない、fetch/pull が publickey で弾かれた、remote-control からリモートの更新を取り込みたい、といった場面で使う。gh-push の取り込み（incoming）版。`/gh-fetch` で明示起動も可。GitHub(gh) 前提。

Audit code quality focusing on deep modules, naming, model hygiene, and architectural clarity. Use after a burst of development, when codebase structure needs assessment, or to make code more agent-navigable.

日本語の概要は準備中です。原文の説明を表示しています。

Knowledge base from the NIST/SEMATECH e-Handbook of Statistical Methods (NIST HB 151) — the practical statistics reference for engineering, metrology, and quality. Use for: exploratory data analysis and the four univariate assumptions (4-plot); measurement process characterization including bias/precision, calibration designs, gauge R&R, and ISO/GUM uncertainty budgets; production process characterization (stability vs. capability); process modeling and regression (LS/WLS/NLS/LOESS); design of experiments (screening, fractional/factorial, response-surface, Taguchi); statistical process control (Shewhart/CUSUM/EWMA charts, capability indices, acceptance sampling); product/process comparisons (hypothesis tests and confidence intervals for 1/2/3+ groups, ANOVA, multiple comparisons); and reliability (lifetime & repair-rate models, accelerated testing, reliability growth). Scope limits: this is applied frequentist statistics for measurement and quality — it does NOT reproduce the per-distribution formula galleries, worked case studies, datasets, plot images, or Dataplot/R code of the original web Handbook (those are described, not copied); it is thin on modern machine learning, Bayesian methods beyond conjugate reliability priors, time-series/forecasting, and Bayesian experimental design.

日本語の概要は準備中です。原文の説明を表示しています。

多模型编排与模型切换策略。基于评分的模型选择、跨 Agent 委派（Gemini、Codex、Ollama）、Advisor 配对、升级触发器、权限矩阵与成本效率优化。

日本語の概要は準備中です。原文の説明を表示しています。

Многомодельная оркестрация и стратегия переключения моделей. Выбор моделей на основе оценки (score), делегирование между агентами (Gemini, Codex, Ollama), связывание с advisor, триггеры эскалации, матрица разрешений и оптимизация затрат.

日本語の概要は準備中です。原文の説明を表示しています。

Enable, disable, or inspect Session Cartographer Turbo Mode, the optional shared warm recall backend for Claude Code and Codex. Use when the user asks about Turbo Mode, faster recall, or the global recall backend.

日本語の概要は準備中です。原文の説明を表示しています。

Cross-vendor code review: this agent reviews the other agent's diff and the other agent reviews this one's, so the final check always comes from a different model. Verdicts are APPROVE or REQUEST_CHANGES with file:line references. Triggers: "/codex-review", "cross-review this diff", "have the other vendor check this".

日本語の概要は準備中です。原文の説明を表示しています。

Master ES6+ features including async/await, destructuring, spread operators, arrow functions, promises, modules, iterators, generators, and functional programming patterns for writing clean, efficient JavaScript code. Use when refactoring legacy code, implementing modern patterns, or optimizing JavaScript applications.

日本語の概要は準備中です。原文の説明を表示しています。