Results for “haiku”

Run fable-mode execution discipline on Claude Haiku. Routes the task to the @fable-worker-haiku agent, whose definition carries the staged loop with tightened verification (no bare "unverified" allowed) and an escalate-don't-improvise rule. Trigger when the user explicitly asks for thorough/systematic handling run cheaply or fast ("fable on haiku", "deep work mode but cheap", "stage this on haiku"). For bulk mechanical work. Do NOT use for tasks needing synthesis — benchmark note: at n=1 the skill's effect on Haiku swung both directions (+25 / −17); route quality-critical work to fable-sonnet instead.

The English summary is being prepared. Showing the original description.

Distill Opus-level reasoning into optimized instructions for Haiku 5.5 (and Sonnet). Generates explicit, procedural prompts with n-shot examples that maximize smaller model performance on a given task. Use when user says "down-skill", "distill for Haiku", "optimize for Haiku", "make this work on Haiku", "generate Haiku instructions", or needs to delegate a task to a smaller model with high reliability.

The English summary is being prepared. Showing the original description.

Compare actual AI session costs with estimates for using only Haiku, Sonnet, or Opus. The skill reports savings amounts and percentages, helping teams assess model selection costs while leaving differences in answer quality and retries outside the calculation.

• Report model selection savings during cost reviews
• Check savings after changing model selection settings
• Use JSON results to enforce a savings threshold in CI

Migrate code and API calls from Claude Sonnet 4.0, Sonnet 4.5, or Opus 4.1 to Opus 4.5. Update platform-specific model identifiers and address reported prompt behavior issues, such as excessive tool use or unnecessary code complexity.

• Updating model identifiers in existing code
• Migrating API settings for a specific platform
• Adjusting prompts that trigger excessive tool use

Choose an agent and model tier using the task description and learned patterns. The skill can explain its recommendation, use compiler-based edits for supported simple changes, and record task outcomes to inform future routing decisions.

• Choose an agent and model before starting a task
• Review the reasoning behind a routing recommendation
• Apply supported code changes through the TypeScript compiler

Add Claude-powered features, such as summarization, to HTML artifacts using a built-in helper. No SDK or API key is needed, and tools defined in the page can run during a request before the final text is returned.

• Adding text summarization to an HTML prototype
• Trying AI responses with conversation history
• Connecting Claude to tools that run within a page

Run fable-mode execution discipline on Claude Fable 5.1 — the top of the escalation ladder and the strongest staged run available. Routes the task to the @fable-orchestrator agent with its model overridden to Fable (`model: "fable"`), which stages the work, delegates artifact production to @fable-worker-sonnet / @fable-worker-haiku, and cold-checks deliverables with @fable-verifier. Trigger when the user explicitly asks for thorough/systematic/"deep work" handling on the strongest model ("fable on fable", "stage this on fable", "deep work mode, fable", "run this on 5.1"), or when a fable-opus run reports the task exceeds Opus's ceiling. Benchmarked 2026-09-02: zero correctness delta on Fable 5.1 at ~1.1× tokens, ~2× wall — a provenance runner, not an accuracy runner. Do NOT use for ordinary single-pass tasks — prefer fable-opus, fable-sonnet, or fable-haiku when the task doesn't need peak reasoning. Fable is plan-included; bulk work still routes down-tier.

The English summary is being prepared. Showing the original description.

Use DeepMind dm-haiku as a JAX neural-network library for transforms, modules, state/RNG, layers, advanced JAX interop, and optional Flax interop.

The English summary is being prepared. Showing the original description.

Run the weak-agent adversarial test harness against docx-cli. Spawns weak exercise agents (Haiku by default, Sonnet to probe, or a local agent harness's pre-produced runs) to perform real document tasks over six scenarios — five editing (MNDA form-fill + font fidelity, invoice table-edit/restructure + logo replace, résumé styling, contract redlining + commenting, contract finalize via accept/reject + comment reply/resolve) and one authoring (T. S. Eliot poetry journal: multi-column, verse, footnotes, links, figure) — renders every result with Word, has opus judge them against ground-truth rubrics, measures each exercise's tool economy, token cost, wall-clock, and correctness (from transcripts for Claude, the exercise.json ledger for the local harness), and synthesizes a prioritized ergonomics report. Use when the user says 'adversarial review', 'test docx-cli with weak agents', 'run the haiku harness', 'weak agent test', or wants to re-run yesterday's adversarial process.

The English summary is being prepared. Showing the original description.

Runs a long swarm of hundreds to thousands of cheap headless coding workers that drive a large failing-test backlog to green. Failures are clustered by signature, each worker fixes one cluster in its own git worktree, a serial no-regression merge gate lands the fixes, the loop resumes across restarts for hours, and it federates across sibling sessions over git branches when one container's CPU is not enough. Use for "fix these thousands of failing tests with a Haiku swarm", "drive the test suite to green with hundreds of agents", "build this compiler or interpreter against a conformance suite", "launch, relaunch or check in on the swarm loop", "the swarm is stuck, idle or rejecting everything", "parallel agents fixing tests keep hitting merge conflicts and regressions", "add sibling containers to the swarm", or "compare models on the tail the swarm could not fix". Ships the harness, a per-project TOML config, a fake-worker selftest, check-in signals and the 2026-10-12 haiku-cc failure catalogue.

The English summary is being prepared. Showing the original description.

Expert guide for intelligent model cascading and routing — complexity-scored task routing from Flash/Haiku to Sonnet/Opus/Astra, dynamic escalation with quality gates, 40-60% token cost reduction while maintaining output quality / Panduan ahli untuk kaskade dan routing model cerdas — routing tugas berbasis skor kompleksitas dari Flash/Haiku ke Sonnet/Opus/Astra, eskalasi dinamis dengan gerbang kualitas, pengurangan biaya token 40-60% dengan kualitas output terjaga.

The English summary is being prepared. Showing the original description.

左脳の解釈を生成する。Gemini CLI や Haiku CLI から実行する。

The English summary is being prepared. Showing the original description.

右脳の解釈を生成する。Gemini CLI や Haiku CLI から実行する。

The English summary is being prepared. Showing the original description.

Measurement-driven model router for Claude Code. Routes Haiku/Sonnet/Opus per task class, logs every routed task with real $ numbers, and A/B tests cheaper tiers before you trust the savings.

The English summary is being prepared. Showing the original description.

Claude API (platform.claude.com) のプロンプトエンジニアリング・評価リファレンス。 prompt engineering, XML タグ, few-shot, system prompts, モデル別プロンプティング（Fable 5.1 / Fable 5 / Opus 5.5 / Opus 5 / Sonnet 5.5 / Sonnet 5 / Haiku 5.5）。 evals, grading（LLM-graded 等）, jailbreak・prompt leak 対策, hallucination 低減, classification, ticket routing, commerce agent 等ユースケース実装。

The English summary is being prepared. Showing the original description.

動画ファイルからアプリの画面遷移リファレンスドキュメントを生成する。video-analyzerスキルのキーフレーム抽出スクリプトでフレームを抽出し、Haikuサブエージェントで並列分析、遷移シーン画像と構造化マークダウン（README.md）を指定ディレクトリに出力する。使用タイミング：動画からアプリのUIフロー・ユーザーストーリーをステップごとにまとめたリファレンスを作成したい場合。

The English summary is being prepared. Showing the original description.

Operating discipline distilled from Claude Fable 5 for sessions and subagents on weaker models (Opus 4.x, Sonnet 5/4.x, Haiku 4.5) doing agentic coding. Load it at the START of any non-trivial task when not running on Fable 5. Операционная дисциплина мышления Fable 5 для менее сильных моделей. Загружай в начале любой нетривиальной задачи (дебаг, рефакторинг, ≥2 файлов, незнакомый код), когда работаешь НЕ на Fable 5 — и указывай сабагенту на Opus/Sonnet/Haiku прочитать этот скилл вместе с нетривиальным заданием. Триггеры — «думай как Fable», «fable-mode», «включи дисциплину», «работай аккуратно», запуск сабагента на слабой модели, повторяющиеся ошибки/переделки в сессии. НЕ нужен для тривиальных правок (rename, опечатка, one-liner) и чисто разговорных ответов.

The English summary is being prepared. Showing the original description.

Operating discipline distilled from Claude Fable 5 for sessions and subagents on weaker models (Opus 4.x, Sonnet 5/4.x, Haiku 4.5) doing agentic coding. Load it at the START of any non-trivial task when not running on Fable 5: debugging, refactoring, changes across 2+ files, or unfamiliar code; and instruct a subagent on Opus/Sonnet/Haiku to read this skill with a non-trivial assignment. Triggers include "think like Fable", "fable-mode", "turn on discipline", "work carefully", dispatching a subagent on a weaker model, and repeated mistakes or rework in the session. NOT needed for trivial edits (rename, typo, one-liner) or purely conversational answers.

The English summary is being prepared. Showing the original description.

Turn Claude Code sessions into reusable working rules with evidence and confidence scores. Keep project-specific habits separate, share general patterns across projects, and group learned rules into skills, commands, or agents for repeated tasks.

• Learning recurring preferences from Claude Code sessions
• Keeping coding conventions separate for each project
• Turning learned rules into skills or commands

Set up Plankton hooks in Claude Code to format and lint files after each edit. Route remaining violations to Claude for fixes, verify the results, and protect linter settings from changes that weaken checks.

• Formatting and linting files after every edit
• Setting up automatic fixes for remaining violations
• Protecting linter configuration from rule weakening

Organize engineering work carried out by AI agents with clear completion criteria, small verifiable tasks, and evaluations before and after implementation. Match model choice to task complexity and track tokens, retries, time, and outcomes to manage cost.

• Planning development work for AI agents
• Splitting a large change into independently verifiable tasks
• Comparing evaluations before and after implementation

Guides implementation and maintenance of Claude applications using language-specific references and official SDKs, the libraries used to call the API. Covers tool integration, streaming responses, model migration, cost reviews, prompt audits, and evaluation workflows.

• Adding Claude API calls to an application
• Building an agent that uses custom tools
• Migrating existing Claude integrations to another model

Locate relevant code in an unfamiliar repository or after a direct search fails. This read-only explorer returns a short list of file paths, verified line ranges, and relevance notes for another agent to use.

• Find starting points in an unfamiliar codebase
• Locate logic spread across multiple files
• Investigate after a direct search fails

Build Claude applications using guidance for the official Anthropic SDK. Choose an integration approach, implement streaming and tool use, and follow workflows for model migration, prompt audits, evaluation, and cost optimization.

• Build a summarization app with Claude API
• Choose how to run an agent with custom tools
• Migrate existing code to a newer Claude model