Communitygithub.com

curiositech/windags-skills

自動化代理技能探索與匹配,實現動態任務路由與能力評估

windags-skills 是什麼?

此技能專注於自主學習系統中的技能發現與驗證,解決「學習目標與達成方法」之間的循環依賴問題。它提供多種架構選擇:對於有明確成功標準的任務,使用單模型快速評估;對於開放式探索任務,採用雙程序架構(系統1快速LLM評估 + 系統2獨立VLM驗證);對於複雜多步驟行為,則透過自上而下的任務分解來發現缺失的中間技能。技能還詳細說明了四種常見失敗模式(橡皮圖章評估、基本動作爆炸、無上下文RAG浪費、庫污染級聯)及其檢測與修復方法,並提供三個完整的工作範例(機器人抽屜開啟、煮咖啡任務分解、RAG環境知識蒸餾)。支援Claude Code、Cursor等AI代理平台,透過清晰的決策樹與驗證策略表,幫助開發者設計穩健的自主學習架構。

相容平台~Claude Code~Codex CLI~Cursor✓Antigravity
npx skills add https://github.com/curiositech/windags-skills/tree/HEAD/skills/agentic-skill-discovery

在你喜歡的 AI 中提問

開啟一個已預先載入此 Agent Skill 的新對話。

說明文件

代理式技能探索

版本: 1.0
領域: 自主學習系統、AI 架構、強化學習

核心概念

自主技能探索解決了雞生蛋蛋生雞的問題:在學習如何達成目標的同時,學習什麼才算成功。系統必須同時扮演學生與老師的角色,這會產生循環依賴,需要在架構上將提案與驗證分離。

決策點

架構選擇決策樹

如果任務有真實成功標準(例如:具有明確目標的導航、操作) → 使用單模型架構搭配快速 LLM 評估 → 成本:低 | 精確度:~75% | 使用案例:預先確定的任務

如果任務需要開放式技能探索(例如:創意任務、探索) → 使用雙程序架構(系統 1/系統 2) → 系統 1:用於訓練循環的快速 LLM 評估 → 系統 2:用於庫接納的獨立 VLM 驗證 → 成本:高 | 精確度:~76% | 使用案例:自主學習

如果需要複雜的多步驟行為 → 自上而下的任務分解 → 從失敗的複雜任務開始 → 分解為子任務 → 學習缺失的先備知識 → 成功率:43.75% 對比 12.50%(自下而上鏈接)

如果要從基本動作建立技能庫 → 僅在基本動作與最終目標完美匹配時使用自下而上的技能鏈接 → 風險:組合爆炸與缺失「中間技能」 → 備援方案:當組合失敗時切換至自上而下

如果當前方法遇到效能瓶頸 → 檢查是否存在循環依賴(同一模型既提案又評估) → 如果檢測到:實施架構分離 → 如果未檢測到:使用成功的技能模式進行 RAG,以進行環境知識蒸餾

驗證策略決策表

場景快速評估(系統 1)慢速評估(系統 2)庫接納規則
訓練循環LLM 生成的成功函數無不適用
技能驗證LLM 評估VLM 驗證兩者皆需通過
庫污染風險高容忍度零容忍度系統 2 覆寫要求
成本限制優化速度優化準確度不對稱成本接受

失敗模式

1. 橡皮圖章評估

症狀: 報告的成功率高但任務完成度低,技能庫快速增長 檢測規則: 如果同一模型生成獎勵並評估成功,則存在循環依賴 根源: LLM 同時扮演玩家與裁判,造成評估偏差 修復: 使用獨立驗證器(例如:視覺任務使用 VLM)實施架構分離

2. 基本動作爆炸癱瘓

症狀: 技能庫龐大但無法解決複雜任務,技能無法有效組合 檢測規則: 如果技能數量增加但複雜任務成功率停滯,則存在「中間技能」問題 根源: 自下而上鏈接無法發現基本動作與複雜目標之間的中間技能 修復: 從期望的複雜行為切換至自上而下的任務分解

3. 無上下文 RAG 浪費

症狀: RAG 檢索已發生但效能未提升,將檢索視為通用範例 檢測規則: 如果 RAG 未改善約束學習(例如:物理理解),則只是昂貴的提示 根源: 缺少環境知識蒸餾機制 修復: 將技能庫結構化為環境賦能性的漸進模型,而不僅僅是範例收集

4. 庫污染級聯

症狀: 隨著技能庫增長,效能隨時間下降,假陽性不斷累積 檢測規則: 如果庫接納使用與訓練循環相同的評估,則污染會級聯 根源: 不良技能導致不良組合;庫錯誤會加乘,不同於訓練雜訊 修復: 庫接納使用昂貴的獨立驗證,訓練循環僅使用廉價評估

5. 人為粒度覆寫

症狀: 技能感覺脫離自然語義邊界,強制抽象層級 檢測規則: 如果手動指定技能粒度與 LLM 提議矛盾,檢查領域匹配性 根源: LLM 從訓練資料中編碼語義任務邊界 修復: 讓系統自由提議技能,觀察發現的粒度,僅根據領域特定需求調整

工作範例

範例 1:使用雙程序驗證的機器人抽屜開啟

場景: 代理必須學習開啟各種抽屜類型,而無預先定義的成功標準。

系統 1(快速)流程:

  1. LLM 生成技能提案:「grasp_drawer_handle」
  2. LLM 建立成功函數:drawer_handle_grasped = gripper_distance < 0.02
  3. RL 訓練使用此函數進行 100 個回合
  4. LLM 評估:「成功!夾爪持續到達把手」
  5. 決策點: 接納入庫?否 - 需要系統 2

系統 2(慢速)流程:

  1. VLM 分析來自 10 個最佳回合的最終影片幀
  2. VLM 裁決:「夾爪觸碰把手但未維持抓握 - 手指未閉合」
  3. 檢測: 系統 1 假陽性被發現
  4. 動作: 拒絕庫接納,改進成功函數
  5. 迭代: 在成功函數中加入抓握壓力要求

專家 vs 新手:

  • 新手遺漏: 信任 LLM 自我評估,接納學習不良的技能
  • 專家發現: 同一模型無法評估其提議的內容;需要獨立驗證

結果: 技能庫維持 76% 的精確度,而無系統 2 時僅為 46%

範例 2:多步驟任務的自上而下任務分解

場景: 「沖泡咖啡」在使用自下而上技能鏈接時失敗,儘管擁有「grasp_cup」、「pour_liquid」等基本動作

自下而上失敗分析:

  • 可用技能:grasp_cup、pour_liquid、move_arm、detect_object
  • 組合嘗試:嘗試了所有 16 種可能的 2 技能鏈
  • 缺失的中間技能:「align_cup_with_spout」、「detect_liquid_level」、「coordinate_bimanual_pouring」
  • 決策點: 自下而上無法發現這些 - 它們不是基本動作的組合

自上而下分解:

  1. 失敗任務:「prepare_coffee」
  2. 分解為:[heat_water] → [grind_beans] → [align_cup] → [pour_water] → [add_beans]
  3. 差距分析:「align_cup」從庫中缺失
  4. 按需學習: 以「grind_beans」的結束狀態作為初始狀態來學習「align_cup」
  5. 鏈接: 每個子任務的成功狀態成為下一個任務的初始條件

權衡分析:

  • 自下而上:12.50% 成功率,探索不必要的技能組合
  • 自上而下:43.75% 成功率,僅學習所需的先備知識
  • 成本: 自上而下需要任務分解能力,自下而上需要詳盡探索

自上而下何時會適得其反:

  • 如果任務分解錯誤,會學習到不正確子任務的技能
  • 如果初始狀態無法可靠設置,鏈接會中斷
  • 緩解措施: 在技能學習前驗證分解,實施狀態設置驗證

範例 3:RAG 環境知識蒸餾

場景: 代理必須在沒有明確物理約束的情況下學習新環境中的操作技能

無 RAG(25% 成功率):

  1. LLM 生成獎勵:reward = distance_to_target
  2. 代理學習接近但不考慮碰撞邊界
  3. 失敗: 不知道在抓取前夾爪必須距離表面 5 公分

有 RAG(46% 成功率):

  1. 檢索先前成功:「reach_cube_A」獎勵函數
  2. 檢測到模式: 所有成功的「reach」都包含 gripper_height > 0.05 約束
  3. LLM 納入:reward = distance_to_target if gripper_height > 0.05 else -10
  4. 知識蒸餾: 隱式學習了環境賦能性

專家 vs 新手:

  • 新手遺漏: 將 RAG 視為範例提供,未識別出約束學習
  • 專家發現: RAG 透過模式識別實現環境知識蒸餾

RAG 何時會適得其反:

  • 如果檢索的模式來自不同環境,則會學到錯誤的約束
  • 如果成功範例過於稀疏,模式不具有統計顯著性
  • 緩解措施: 按環境相似性過濾檢索,要求模式提取的最小樣本量

品質關卡

  • 提案與驗證使用不同的模型能力(例如:LLM 提案,VLM 驗證)
  • 在保留測試集上測量,快速評估精確度 >45%,慢速評估精確度 >75%
  • 庫接納需要獨立驗證,而不僅是訓練循環成功
  • 複雜任務分解產生 3-7 個有意義的子任務,且具有清晰的依賴排序
  • RAG 檢索改善約束學習的程度 > 無 RAG 基準線的 15%
  • 技能粒度分析顯示語義聚類,而非參數列舉
  • 系統能夠檢測並從循環依賴失敗模式中恢復
  • 透過獨立驗證,庫污染率 <25%
  • 自上而下分解在複雜任務上的表現優於自下而上鏈接 >2 倍
  • 環境知識蒸餾展示從成功模式中學習約束

非適用範圍

請勿將此技能用於:

  • 具有已知成功標準的預先確定任務 → 改用帶有手動設計獎勵的標準 RL
  • 單步驟操作任務 → 改用帶有示範資料的監督式學習
  • 需要即時效能的任務 → 雙程序驗證太慢;改用快取技能庫
  • 缺乏視覺回饋的領域 → VLM 驗證不可用;改為設計領域特定的獨立驗證器
  • 資源受限環境 → 系統 2 驗證成本高昂;接受較低精確度或使用分層驗證

改委派給以下技能:

  • 標準獎勵工程 → 用於具有清晰成功指標的任務
  • 模仿學習 → 用於具有大量示範資料的任務
  • 分層 RL → 用於具有已知技能分解的任務
  • 元學習 → 用於快速適應新任務分佈
  • 憲法 AI → 用於需要價值對齊與安全約束的任務

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

curiositech/windags-skills

Expert in 2000s-era music visualization (Milkdrop, AVS, Geiss) and modern WebGL implementations. Specializes in Butterchurn integration, Web Audio API AnalyserNode FFT data, GLSL shaders for audio-reactive visuals, and psychedelic generative art. Activate on "Milkdrop", "music visualization", "WebGL visualizer", "Butterchurn", "audio reactive", "FFT visualization", "spectrum analyzer". NOT for simple bar charts/waveforms (use basic canvas), video editing, or non-audio visuals.

curiositech/windags-skills

Expert legal research agent for finding and scraping expungement data state by state. Knows authoritative sources, URL patterns, Firecrawl configuration, and 2026 legal landscape.

curiositech/windags-skills

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

curiositech/windags-skills

Implement WCAG 2.2 AA/AAA compliance with automated testing, keyboard navigation, screen reader support, and focus management. Activate on: accessibility audit, WCAG compliance, keyboard navigation, screen reader, aria attributes, axe-core, focus trap. NOT for: design-level accessibility review (use design-accessibility-auditor), color contrast only (use css-in-js-architect).

curiositech/windags-skills

Time-blind friendly planning, executive function support, and daily structure for ADHD brains. Specializes in realistic time estimation, dopamine-aware task design, and building systems that actually work for neurodivergent minds.

curiositech/windags-skills

Designs digital experiences for ADHD brains using neuroscience research and UX principles. Expert in reducing cognitive load, time blindness solutions, dopamine-driven engagement, and compassionate design patterns. Activate on 'ADHD design', 'cognitive load', 'accessibility', 'neurodivergent UX', 'time blindness', 'dopamine-driven', 'executive function'. NOT for general accessibility (WCAG only), neurotypical UX design, or simple UI styling without ADHD context.

curiositech/windags-skills

>- Apply crisis decision-making research to agent routing, uncertainty triage, and coordination failure analysis in time-pressured systems. Use when diagnosing handoff failures, analytical paralysis, or expert judgment under incomplete information. NOT for routine coding, simple CRUD design, or static single-agent tasks with complete information.

curiositech/windags-skills

Extend and modify the admin dashboard, developer portal, and operations console. Use when adding new admin tabs, metrics, monitoring features, or internal tools. Activates for dashboard development, analytics, user management, and internal tooling.

curiositech/windags-skills

Conversation patterns and interaction protocols for multi-agent systems. Covers request/response, pub/sub, blackboard, delegation chains, debate, critique, consensus, fan-out/fan-in, supervisor-worker, and peer negotiation. Deep analysis of AutoGen conversation patterns, CrewAI delegation, LangGraph state passing, and FIPA-ACL performatives. Teaches how to design what agents say to each other and in what order. Activate on: "agent conversation", "agent protocol", "multi-agent debate", "agent delegation", "supervisor worker pattern", "agent voting", "consensus protocol", "fan-out fan-in", "agent negotiation", "blackboard pattern", "agent dialogue", "conversation topology", "agent handoff". NOT for: wire format or serialization (use agent-interchange-formats), orchestration infrastructure (use agentic-infrastructure-2026), single agent behavior (use agentic-patterns).

curiositech/windags-skills

Meta-agent for creating new custom agents, skills, and MCP integrations. Expert in agent design, MCP development, skill architecture, and rapid prototyping. Activate on 'create agent', 'new skill', 'MCP server', 'custom tool', 'agent design'. NOT for using existing agents (invoke them directly), general coding (use language-specific skills), or infrastructure setup (use deployment-engineer).

curiositech/windags-skills

AI-powered calendar management and agent-based scheduling coordination. Covers calendar APIs (Google Calendar, CalDAV/iCal), AI scheduling assistants (Reclaim, Clockwise, Motion, Cal.com), building custom calendar agents with MCP, multi-calendar merging, timezone management, focus block protection, meeting fatigue detection, and agent-to-agent meeting negotiation protocols. Activate on: "calendar agent", "AI scheduling", "calendar coordination", "meeting scheduling", "calendar API", "focus time protection", "calendar optimization", "Google Calendar MCP", "Reclaim", "Clockwise", "Motion", "Cal.com", "smart scheduling", "calendar-aware agent", "timezone scheduling", "agent negotiation meetings". NOT for: manual calendar UI component design (use form-validation-architect), project management scheduling or Gantt charts (use project-management-guru-adhd), general time-tracking or pomodoro apps (use adhd-daily-planner for time-awareness), building the agent itself from scratch (use agent-creator).

curiositech/windags-skills

Build and adopt production AI agent infrastructure in 2026. Covers framework selection (LangGraph, CrewAI, AutoGen, MCP), orchestration patterns, evaluation, observability, memory systems, and tool use. Also covers the SOCIAL dimension: how to sell agent infrastructure internally, change management, measuring ROI, building trust in autonomous systems, and scaling adoption across teams. Activate on: "agent infrastructure", "agent framework comparison", "which agent framework", "sell AI tools internally", "agent adoption", "agent observability", "agent evaluation", "MCP architecture", "agentic mesh", "enterprise AI agents", "AI change management", "agent ROI". NOT for: building specific agents (use ai-engineer), designing agent behavior patterns (use agentic-patterns), prompt tuning (use prompt-engineer).

curiositech/windags-skills

Fundamental patterns for effective agentic behavior. Teaches decomposition, tool orchestration, error recovery, context management, quality self-assessment, and knowing when to stop. Model-agnostic principles that make any agent more effective regardless of domain. Activate on: "how should I structure this agent", "agentic workflow", "agent patterns", "multi-step task", "tool orchestration", "/agentic-patterns", "decompose this", "agent best practices", "chain of actions", "when should the agent stop", "agent loop design". NOT for: creating agent infrastructure (use agent-creator), building DAGs (use windags-architect), specific tool implementation.

curiositech/windags-skills

Cryptographic security for agentic systems — zero-trust agent networking, signed message envelopes (JWS/JWE), capability-based security (ocaps), Merkle tree audit trails, WASM sandboxing, and formal verification. Covers CLI dev tool security, mTLS between agents, permission boundaries (least privilege for AI agents), and supply chain security for skills/plugins. Activate on: "agent security", "zero trust agents", "secure agent communication", "capability-based security", "ocap", "signed messages between agents", "agent audit trail", "sandbox agent execution", "agent permissions", "mTLS agents", "cryptographic verification", "agent supply chain", "OWASP agentic", "prove agent did X", "tamper-proof agent logs". NOT for: application-level SAST scanning (use security-auditor), network firewall rules (use infrastructure), SOC2/HIPAA compliance (organizational), or prompt injection defense (use prompt-engineer).

curiositech/windags-skills

Data structures and serialization formats for agent-to-agent communication. Covers message envelopes, structured output schemas, capability declarations, task handoff payloads, error/retry signaling, and context windows as data structures. Deep comparison of A2A protocol, MCP, OpenAI function calling, and LangChain message types. Teaches when to use rigid schemas vs free-form with validation, typed vs untyped, streaming vs batch. Activate on: "agent message format", "agent communication schema", "agent-to-agent protocol", "A2A protocol", "MCP message format", "structured output for agents", "agent interop", "interchange format", "agent serialization", "task handoff format", "capability declaration". NOT for: what agents say to each other (use agent-conversation-protocols), orchestration topology (use multi-agent-coordination), building agent infrastructure (use agentic-infrastructure-2026).

curiositech/windags-skills

Logic-based agent programming language implementing BDI architecture for practical autonomous agent development

curiositech/windags-skills

>- Design AgentSpeak(L)-style BDI agents with context-guarded plans, selection functions, and intention stacks. Use for interruptible autonomy, agent policy, and multi-agent orchestration in dynamic environments. NOT for simple rule engines, static planners, or centralized workflows.

curiositech/windags-skills

Foundational concurrent computation model where actors communicate exclusively through asynchronous message passing

curiositech/windags-skills

Build production-ready LLM applications, advanced RAG systems, and intelligent agents. Implements vector search, multimodal AI, agent orchestration, and enterprise AI integrations. Use PROACTIVELY for LLM features, chatbots, AI agents, or AI-powered applications.

curiositech/windags-skills

license: Apache-2.0 NOT for unrelated tasks outside this domain.

相關技能