「tuning」の検索結果

Use when operating ProjectClownfish cluster workflows in this repo: reviewing GitHub Actions run artifacts, tuning worker/applicator guardrails, importing ghcrawl clusters, dispatching canaries, scaling batches, or deciding whether autonomous write-mode cluster cleanup is safe to ramp.

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。

Diagnoses and fixes Claude behavioral issues in system prompts and Projects using tested countermeasure templates. Use when "Claude is too verbose," "keeps hedging," "agrees with everything," "claims it did something it didn't," "won't use a tool," "fix Claude's output," "tune Claude's behavior," "Claude ignores my preferences," "adds unsolicited disclaimers," "uses too many lists." Covers ten tendencies: agreeableness (output-content + persistent-preference), hedging, verbosity, list overuse, fabricated precision, over-exploration, tool miscalibration (over- and under-triggering), LaTeX defaulting, editorial drift, self-referential fabrication. Also use when auditing a system prompt for behavioral calibration or recalibrating a pre-4.7 prompt, and when users describe recurring output problems without naming a tendency. Do NOT use for scoring a prompt's overall quality — use rootnode-prompt-validation if available.

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。

Generalised autonomous optimisation loop — soft RLVR for any artifact a user can measure. Use this skill whenever a user wants to iteratively improve an artifact — code, prompts, documents, configs, designs, content — by running structured experiments, evaluating results against a multi-dimensional rubric, and learning from each attempt. Triggers include: "optimise this", "keep improving until it's good", "run experiments on", "autoresearch", "iterate on this overnight", "try different approaches and pick the best", or any request implying repeated evaluate-and-improve cycles. Also use when the user wants to improve a system prompt, a data pipeline, a writing style, or any artifact where quality can be decomposed into measurable tracks. For inference optimisation tasks (model latency, throughput, quantization, GPU deployment), a* delegates the low-level tuning to AITune while maintaining quality tracking and learning.

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。

Implement and execute approved research Experiment, Tuning, or Design Cards with reproducible logs, metrics, environment, and artifacts. Use for bounded coding, debugging, and runs; not for changing claims, priorities, budgets, or the Research Spine.

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。

Apply collective intelligence coordination patterns — stigmergy, local rules, and quorum sensing — to organize distributed systems, teams, or workflows without centralized control. Covers signal design, agent autonomy boundaries, emergent behavior cultivation, and feedback loop tuning. Use when designing distributed systems without a coordination bottleneck, organizing teams that must self-coordinate, building event-driven architectures with shared state communication, or replacing fragile centralized orchestration with resilient emergent coordination.

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。

Systematically identify whether a GPU kernel is compute-bound, memory-bound, or latency-bound using roofline analysis, occupancy calculations, compute/load ratio per tile, and SASS instruction inspection. Produces a decision matrix for optimization strategy selection (cp.async, warp interleaving, tiling, double-buffering, or CuAssembler hand-tuning).

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 implementing pool/slab/arena allocators, tuning jemalloc/mimalloc/tcmalloc, writing a Rust GlobalAlloc, or benchmarking allocator performance and fragmentation.

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 tuning or profiling native code on Apple M-series Macs: unified memory, 16 KiB pages, Accelerate and Metal for matrix work, xctrace and leaks, Rosetta 2, or sysctl hardware 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。

Use when writing or restructuring code, before adding a helper, wrapper, config key, or dependency, or when the user asks for minimal or DRY code. Not for performance tuning: use optimize.

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。

End-to-end system for creating supervised fine-tuning datasets from books and training style-transfer models. Covers text extraction, intelligent segmentation, synthetic instruction generation, Tinker-compatible output, LoRA training, and validation.

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 SurrealDB 3 skill. Use when working with SurrealDB, SurrealQL queries, multi-model data modeling (document, graph, vector, time-series, geospatial), schema design, graph traversal, vector search, security and permissions, deployment and operations, performance tuning, SDK integration, Surrealism WASM extensions, SurrealMCP, Surrealist IDE, Surreal-Sync migrations, SurrealFS, SurrealKit, or ecosystem integrations such as n8n and CodeMirror.

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 SurrealDB 3 architect and developer skill (tracks v3.1.4+). SurrealQL mastery, multi-model data modeling (document, graph, vector, time-series, geospatial), schema design, security, deployment, performance tuning, SDK integration (JS, Python, Go, Rust, Java, .NET, C, PHP, Swift, Kotlin, Ruby), Surrealism WASM extensions, SurrealML scope coverage (preview), built-in MCP (v3.1+) and standalone SurrealMCP, LangChain Python integration, editor tooling (LSP v0.1.6, tree-sitter, CodeMirror v1.0.6, VS Code/JetBrains/Neovim/Zed), and ecosystem integrations (Surrealist v3.9, Surreal-Sync, SurrealFS, SurrealKit v0.7, n8n, Agent Skills, setup-surreal). Universal skill for 30+ AI agents.

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。

量化专家经验调优库：按「L1 通用专家调优意见 → L2 结构化专家调优意见 → L3 模型专属专家调优意见」三级递进， 输出量化精度调优的完整手段（离群值抑制 / 量化方法·粒度·对称性选型 / 校准集调整 / 敏感层回退 / 模型专属策略）， 每条结论附专家原因分析与专家意见可信度等级。 本 Skill 是专家经验库，负责说明可选离群值抑制算法并询问用户选择哪些算法，以及回答「怎么调、哪些层需要回退及为什么」；不运行离群值抑制 processor，不修改 YAML、不执行量化、不做 EP 检查 / 服务化 / 任务评测。

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。

执行模型量化。通过 msmodelslim quant 依据 Practice YAML 对模型进行量化。

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。

执行模型测评（统一入口）。LLM/VLM 走主流程 scripts/run_evaluation.py + Evaluation YAML；model_family=dit 走 DiT 扩展节（references/dit/evaluate_workflow.md，vbench.py 批量推理产出视频）。

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。

Customize any Cowork skill or workflow to match your specific industry, company, and style. For tuning default skill outputs to your voice, when adapting generic workflows to a specific industry, or during a skill quality review. Use when the user says "customize this skill", "output is too generic", "make this work for my industry", "personalize this workflow", "results do not match my style", or "how do I get better results from Cowork".

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。

MATLAB R2026a agentic workflow router. Use this whenever the user asks for MATLAB, Simulink, .m/.mlx/.slx work, toolbox-driven engineering, paper reproduction, simulation, code generation, data analysis, control, robotics, signal/image/AI workflows, or wants an end-to-end automated MATLAB task with validation. For Simulink tasks on this machine, default to a visible workflow: open MATLAB/Simulink windows first so the user can watch modeling, simulation, tuning, and plotting progress.

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。

Simulate supply chain and adversarial machine learning attacks by injecting poisoned data or targeted backdoors into training and fine-tuning datasets. Use this skill when assessing the integrity controls of MLOps pipelines or evaluating the resilience of AI models against highly targeted, stealthy manipulation intended to alter model behavior on specific triggers.

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。

Identify and simulate Data Poisoning attacks aimed at degrading or skewing an AI model's accuracy. This skill focuses on Adversarial Machine Learning concepts where attackers inject malicious or mislabelled data points into training or fine-tuning datasets (e.g., feedback loops) to bias the AI.

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。

Advanced techniques for bypassing LLM safety filters, instruction tuning, and system prompt restrictions using specialized linguistic constructs, hypothetical scenarios, and persona adoption.

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。

Together AI SDK patterns for TypeScript — client setup, chat completions, streaming, structured output, function calling, embeddings, image generation, fine-tuning, and OpenAI-compatible endpoints

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。

Uses YARA to classify and triage samples at scale: applying rule sets, reading matches and string offsets, tuning for false positives, and organizing rules for malware family identification. Activates for requests to scan files with YARA, apply YARA rules, or classify samples by signature.

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。

Converts a successful threat hunt into a durable detection by extracting the discriminating logic, defining data sources and thresholds, and emitting a Sigma rule plus a test and tuning plan. Activates for requests to operationalize a hunt, turn a finding into a detection rule, or productionize hunt logic as a Sigma detection.

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 design spec for small changes — tuning adjustments, minor mechanics, balance tweaks. Skips full GDD authoring when a system GDD already exists or the change is too small to warrant one. Produces a Quick Design Spec that embeds directly into story files.

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。