「lora」の検索結果

Autonomously improve a real artifact (code, training recipe, agent harness, data pipeline, prompt) against an objective and an evaluator, using Hypothesis Tree Refinement (HTR) from the Arbor paper. Use this whenever someone wants to iteratively optimize something over many experiments without overfitting — e.g. "get my model's eval score up", "improve this agent/harness", "tune this pipeline", "beat the baseline on this benchmark", "run a search over approaches and keep the best", "do an MLE-bench / Kaggle-style optimization", or any long-horizon "make this artifact better and don't just memorize the dev set" task. Trigger it even when the user doesn't say "Arbor" or "hypothesis tree" but describes repeated experiment-and-evaluate loops, branching exploration of competing ideas, or worries about a dev/test gap. Runs Claude itself as the coordinator with subagent executors in isolated git worktrees; for the standalone `arbor` CLI tool see references/arbor-upstream.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。

Workflow Phase 3 — Research and Solution Exploration. Surfaces 2–4 candidate solution approaches with concrete tradeoffs, grounded in current library docs (fetched through the provider named in `.claude/docs-provider.json`) — never in training-data recall. Output lives at `docs/research/<slug>.md`. Candidate ranking and the memo execute in main context; doc/source gathering MAY be delegated to read-only advisory subagents (seed.md §4.2-A).

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 ane for ALL file exploration, discovery, reading, and editing unless the user states otherwise. `ane exec <file> --chord "<chord>"` handles reads (Yank), searches (List), and edits (Change, Delete, Replace, Append, Prepend, Insert). Prefer ane over builtin file tools and shell utilities like grep, sed, and cat.

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。

Multi-agent collaboration plugin that spawns N parallel subagents competing on the same task via git worktree isolation. Agents work independently, results are evaluated by metric or LLM judge, and the best branch is merged. Use when: user wants multiple approaches tried in parallel — code optimization, content variation, research exploration, or any task that benefits from parallel competition. Requires: a git repo.

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。

Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO).

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。

Exploratory QA of web apps: find bugs, evidence, reports.

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。

Generate beautiful, on-brand HTML artifacts — reports, diagrams, diff reviews, slide decks, plans, recaps, dashboards. Loads on creation intents ("create a plan", "make me an implementation plan", "generate a diagram", "build a slide deck", "draft a recap", "visualize this", "render this as HTML", "make me a status update") and exploration intents ("show me three layouts", "explore document shapes", "compare visual treatments", "explore a new design system", "compare replacement brand directions") and iteration intents ("make this beautiful", "polish this report", "make this less generic", "design this", "theme this", "use my brand", "iterate on this design", "review this artifact", "make this bolder", "make this quieter", "simplify this design", "animate this", "improve the look of", "give this a refresh"). Captures brand from the user's codebase, asking, or live site, then applies it to every artifact.

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。

Expert-level planetary science covering solar system formation, planetary interiors, atmospheres, surface processes, astrobiology, and planetary exploration missions.

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。

Expert-level geophysics covering seismic methods, gravity and magnetics, electromagnetic methods, heat flow, Earth structure, and exploration geophysics.

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。

Fine-tune LLMs and VLMs using Unsloth on HF Jobs (Hugging Face on-demand cloud GPUs). Use when users want to fine-tune language models, train VLMs (Vision Language Models), do continued pretraining, domain adaptation, or run UV scripts on HF Jobs. Triggers on requests involving Unsloth training, HF Jobs GPU training, Qwen3-VL fine-tuning, Gemma VLM training, or LoRA fine-tuning on cloud GPUs.

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。

Você DEVE usar isto antes de qualquer trabalho criativo - criar features, construir componentes, adicionar funcionalidade ou modificar comportamento. Explora a intenção do usuário, requisitos e design antes da implementação.

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。

Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a PPT/PPTX to web, or create slides for a talk/pitch. Helps non-designers discover their aesthetic through visual exploration rather than abstract choices.

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。

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.

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。

GStack design-html method for producing or planning quick HTML prototypes, design boards, or rendered visual explorations for UI review.

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。

Track Two / internal innovation system for running small, fast experiments outside the core business. Runs in three modes — Diagnostic for designing the full Protoloop from scratch (intake channels, loop structure, team, exit paths), Review for running the monthly heartbeat meeting (loop-by-loop decisions, forward-loading, portfolio health), and Alert Triggers for defining tripwires that protect, feed, and maintain momentum in the system. Updates company-context.md after every run. Triggers on questions about innovation pipelines, intrapreneurship, Track Two strategy, experimentation systems, handling ideas that don't fit the core business, or building systematic exploration alongside exploitation. Based on 20 years of practice by Eterdis (eterdis.com).

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。

Lightweight multi-agent research orchestration using one coordinator, three domain leads, and three lightweight research agents. Use for medium-size research, market scans, competitor comparisons, prior-art style exploration, report planning, and any task that benefits from parallel domain decomposition without running a full high-cost research lab.

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。

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to ...

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。

Delegate coding tasks to Codex, Claude Code, or Pi agents via background process. Use when: (1) building/creating new features or apps, (2) reviewing PRs (spawn in temp dir), (3) refactoring large codebases, (4) iterative coding that needs file exploration. NOT for: simple one-liner fixes (just edit), reading code (use read tool), thread-bound ACP harness requests in chat (for example spawn/run Codex or Claude Code in a Discord thread; use sessions_spawn with runtime:"acp"), or any work in ~/clawd workspace (never spawn agents here). Claude Code: use --print --permission-mode bypassPermissions (no PTY). Codex/Pi/OpenCode: pty:true required.

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。

Exploratory data-migration validation and QA ideation workflow. Use when you need to discover hidden risks, generate new validation angles, and converge mapping specs/validation reports into a prioritized QA backlog for CRM migrations. Version 2.0, updated 2025-12-28.

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。

Comprehensive data science workflow skill for analyzing tabular and time series data. Use when conducting exploratory data analysis, building predictive models, performing feature engineering, or generating data-driven insights. Supports end-to-end analysis from data understanding through model deployment, following industry best practices for regression, classification, time series forecasting, and statistical analysis.

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 designing, reviewing, modifying, migrating, optimizing, securing, operating, troubleshooting, or decommissioning data architectures across data model exploration, SQL, ISO SQL conformance, NoSQL, vector, graph, document, time-series, NewSQL, search, in-memory, SQLite, CDC/Debezium, DBMS storage/query/recovery internals, database incident prevention, hardware acceleration, cache, big-data, lakehouse, and warehouse systems.

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。

ego-browser (ego-lite) is a Chromium-based browser that gives AI agents a CLI-accessible Node.js runtime for driving a real browser. Use this skill whenever the user needs to interact with a website opening pages, filling forms, clicking buttons, taking screenshots, extracting page data, testing web apps, logging into sites, automating browser operations, or any other browser automation task. Triggers include requests to "open a website", "visit a URL", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "extract content from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also used for exploratory testing, dogfooding, QA, bug hunting, or reviewing app quality. Prefer ego-browser over any built-in browser automation, web fetch, or other web tools.

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。

Plan a refactor as a sequence of tiny safe commits via interview and codebase exploration. Use when restructuring working code, or when the user says 'plan a refactor'.

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。