Configure with gradient clipping helper operations. Auto-activating skill for ML Training. Triggers on: gradient clipping helper, gradient clipping helper Part of the ML Training skill category. Use when working with gradient clipping helper functionality. Trigger with phrases like "gradient clipping helper", "gradient helper", "gradient".
Optimize tailwind class optimizer operations. Auto-activating skill for Frontend Development. Triggers on: tailwind class optimizer, tailwind class optimizer Part of the Frontend Development skill category. Use when working with tailwind class optimizer functionality. Trigger with phrases like "tailwind class optimizer", "tailwind optimizer", "tailwind".
Brief description of what this skill does and when the model should activate it. Use when [describe the user's intent or situation]. Trigger with "example phrase", "another trigger", "/example-skill".
Intelligent AI model router that automatically switches between two configured models (local for simple tasks, cloud for complex ones). Detects local models (Ollama, LM Studio) automatically, routes based on task complexity and privacy. Use when users ask to "switch model", "use local/cloud model", or mention API keys/passwords (triggers privacy mode). Trigger on: sensitive data detection (forces local), complex tasks like "design architecture" (uses cloud), or configuration requests.
Intelligent AI model router that automatically switches between two configured models (local for simple tasks, cloud for complex ones). Detects local models (Ollama, LM Studio) automatically, routes based on task complexity and privacy. Use when users ask to "switch model", "use local/cloud model", or mention API keys/passwords (triggers privacy mode). Trigger on: sensitive data detection (forces local), complex tasks like "design architecture" (uses cloud), or configuration requests.
Intelligent AI model router that automatically switches between two configured models (local for simple tasks, cloud for complex ones). Detects local models (Ollama, LM Studio) automatically, routes based on task complexity and privacy. Use when users ask to "switch model", "use local/cloud model", or mention API keys/passwords (triggers privacy mode). Trigger on: sensitive data detection (forces local), complex tasks like "design architecture" (uses cloud), or configuration requests.
本skill基于东方财富妙想搜索能力，基于金融场景进行信源智能筛选，用于获取涉及时效性信息或特定事件信息的任务，包括新闻、公告、研报、政策、交易规则、具体事件、各种影响分析、以及需要检索外部数据的非常识信息等。避免AI在搜索金融场景信息时，参考到非权威、及过时的信息。
Capture a correction or lesson as a persistent learning rule with category, mistake, and correction. Stores, categorises, and retrieves rules for future sessions. Use after mistakes or when the user says "remember this", "don't forget", "note this", or "learn from this".
VeOmni-specific checklist for feature development and refactoring. Covers impact analysis across modalities, trainer hierarchy, data pipeline, and distributed code. Use before implementing any non-trivial change. For model-specific or ops-specific work, use veomni-new-model or veomni-new-op instead. Trigger: 'add feature', 'implement', 'refactor', 'reorganize', 'new capability'.
Use this skill when adding support for a new model to VeOmni. Covers the full lifecycle: analyzing the HuggingFace model, creating model patches, defining parallel plans, writing configs, integrating with the trainer, and testing. Trigger: 'add model', 'support new model', 'integrate <model_name>', 'new model support'.
Iterative skill tuning via execute-evaluate-improve feedback loop. Uses ccw cli Claude to execute skill, Gemini to evaluate quality, and Agent to apply improvements. Iterates until quality threshold or max iterations. Triggers on "skill iter tune", "iterative skill tuning", "tune skill".
Meta-skill for generating team skills following the v4 architecture pattern. Produces complete skill packages with SKILL.md router, coordinator, worker roles, specs, and templates. Triggers on "team-designer", "design team".
Unified team skill for plan-and-execute pipeline. Pure router — coordinator always. Beat model is coordinator-only in monitor.md. Triggers on "team planex".
Chain-loaded review cycle. 3-mode routing for session, module, or fix review.
Chain-loaded spec generation. 8-phase with quality gate loop and auto-fix.
Optimizes AI skills for activation, clarity, and cross-model reliability. Use when creating or editing skill packs, diagnosing weak skill uptake, reducing regressions, tuning instruction salience, improving examples, shrinking context cost, or setting benchmark/release gates for skills. Trigger terms: skill optimization, activation gap, benchmark skill, with/without skill delta, regression, context budget, prompt salience.
Steps for creating or modifying a Django model
Baidu FaMou algorithm skills for efficient algorithm self-evolution. Provides experiment management and visualization capabilities to help optimize complex algorithms. Use when user needs algorithm optimization or experiment management.
Use this skill whenever creating a new lesson, choosing a lesson pattern, or validating lesson structure. Trigger on: 'create a lesson', 'lesson recipe', 'lesson pattern', 'terminal-only lesson', 'code editing lesson', 'database lesson', 'console lesson', 'lesson sequence', 'multi-step lesson', 'validate lesson', 'check lesson', or asking how to structure a Rails lesson — even without mentioning recipes. Provides five tested blueprints (terminal-only, code-editing, database, full-app, console/IRB) with correct frontmatter, directory layouts, content examples, and a post-creation validation checklist. Do NOT create lessons without this skill — incorrect structure causes silent failures. Do NOT use for frontmatter reference (use tutorial-lesson-config) or file organization (use rails-file-management).
Use this skill whenever checking if a Rails feature or gem works in WASM, or understanding what's real vs. conceptual in lessons. Trigger on: 'does X work in WASM', 'can I use this gem', 'gem compatibility', 'WASM limitation', 'what works', 'supported features', 'PGLite', 'threading', 'Net::HTTP', 'ActionCable', 'background jobs', 'can I teach X', 'bundle install in WASM', 'conceptual vs real', or any Rails capability question — even without mentioning WASM. Authoritative compatibility matrix: which features work, which are shimmed, which are impossible, plus gem tiers, PGLite behavior, boot timing, and conceptual-vs-real operations. General Rails knowledge is insufficient here. Do NOT use for file organization (use rails-file-management) or frontmatter (use tutorial-lesson-config).
Use this skill whenever configuring lesson frontmatter or checking inheritance rules. Trigger on: 'frontmatter', 'prepareCommands', 'mainCommand', 'editor config', 'terminal config', 'preview config', 'configuration inheritance', 'focus file', 'i18n', 'allowEdits', 'previews', 'autoReload', 'scope', 'defaults', 'what inherits', 'does X inherit', 'invalid combination', 'constraint', or any YAML frontmatter question — even without mentioning configuration. Authoritative reference for all frontmatter options, inheritance rules, defaults, and invalid-combination constraints with Rails-specific patterns. Do NOT guess frontmatter without this skill. Do NOT use for content hierarchy (use tutorial-content-structure) or file organization (use rails-file-management).
Benchmark inference performance for a specific model
Get list of currently downloaded models
