Gather phase context through adaptive questioning before planning. Use --auto to skip interactive questions (Claude picks recommended defaults). Use --chain for interactive disc...

anti-sycophancy

职位市场分析。触发场景：用户提供职位名称和地区，要求分析职位市场供需、薪资水平、技能趋势。

简历诊断。触发场景：用户要求诊断简历质量；用户想优化简历; 用户问我的简历有什么问题。

简历解析。触发场景：用户上传简历文件要求解析、提取结构化信息。

Master the art of spawning and managing sub-agents. Write prompts that actually work, track running agents, and learn from every outcome. Part of the Hal Stack 🦞

基于文本匹配的快速路由引擎，为简单任务提供零 LLM 决策的快速路径。支持中英文混合输入，自动匹配技能/工具，低置信度时降级到 LLM 语义路由。

reprompter

A comprehensive skill to protect against prompt injection attacks when processing external content such as websites, emails, documents, and other untrusted input sources.

Agent Lightning를 사용하여 AI 에이전트를 자동으로 최적화하는 방법을 제공합니다.

<skill>

Activate a beta feature using an activation code

This skill provides knowledge about Claude tools (Prompts, Skills, Subagents, Commands, and MCPs). Use this skill when knowledge about Claude tools is needed, when the user mentions Claude tools or when the user needs help identifying the best Claude tool to use/create for a task. Common questions include 'Should I create a Claude tool for this use-case?', 'What type of Claude tool should I create for this use-case?.

Reads config.json and prompts user for application parameters. Merges configuration with defaults for project generation.

Expert knowledge on creating slash commands for Claude Code. Use when designing or creating slash command .md files, understanding command arguments, or implementing bash execution.

Delphi launches multiple oracle agents in parallel with the **same prompt** to discover divergent paths of exploration. By giving identical starting points, each oracle independently explores the prob

Anthropic의 Best Practice를 기반으로 고성능 프롬프트를 설계하고 최적화합니다.

Run diagnostic checks on GSD installation and report status

'Ready-to-use prompts for comprehensive .NET framework upgrade analysis and execution'

'Structured Autonomy Planning Prompt'

task = $ARGUMENTS

AI工程师角色。关键词：AI/LLM/智能体/Prompt/上下文工程/多智能体/工作流/RAG/Function Call/效果评测。激活后设计提示词、构建智能体调用链、评测输出质量。

Create detailed phase plan (PLAN.md) with verification loop

Plugins can store user-configurable settings and state in `.claude/plugin-name.local.md` files within the project directory. This pattern uses YAML frontmatter for structured configuration and markdow

Autonomous task execution via the codex CLI. Runs non-interactively. Progress streams to stderr; final result on stdout.

Spawns multiple AI coding agents to work on related GitHub issues concurrently using git worktrees. Use when breaking down a large feature into multiple issues, running parallel agents with --print flag, or managing wave-based execution of related tasks.

robius-widget-patterns

Activate autonomous Ralph Wiggum loop mode for iterative task completion. Use when you have a well-defined task with clear completion criteria that benefits from persistent, autonomous execution.

این مهارت برای ایجاد توابع پیشرفت (progress callback) در کتابخانه Pyrogram استفاده می‌شود تا هنگام آپلود یا دانلود فایل، با بررسی تغییر وضعیت و محدود کردن طول متن، از خطاهای `MESSAGE_NOT_MODIFIED` و `MESSAGE_TOO_LONG` جلوگیری شود.

Создает PHP-скрипт для парсинга MHTML-файлов без сторонних библиотек, извлекает заданные поля (название, просмотры, ссылка) и выводит их в HTML-таблицу со стилями Bootstrap.

使用python-docx库编写Python脚本，递归遍历指定文件夹（包含子文件夹）中的所有Word文档，批量提取其中的段落文本和表格内容。

从OCR识别后的医疗票据文本中提取日期、医生姓名、病人姓名、诊断和总消费，并进行文本矫正，输出JSON格式。

使用PyMuPDF库将现有PDF文件的所有页面内容复制并重绘到新的PDF文件中，严禁使用渲染成图片（栅格化）的方式处理，以保留文本的可选性和矢量特性。

创建一个覆盖全屏的透明蒙版，监听触摸事件，在用户触摸时打开指定 URL 并自动删除蒙版。

实现一个支持做多和做空操作的股票交易强化学习环境，采用未实现盈亏作为奖励机制，并在每轮结束后异步保存K线图和资金曲线。

Revises academic text to enforce active voice, clarity, and brevity. Optionally integrates transition words and breaks down complex sentences to enhance flow and readability.

A coding assistant that first outlines a detailed plan in pseudocode, then outputs the final code in a single code block with minimal additional text.

Develops comprehensive security plans for high-risk corporate meetings or board events, focusing on venue hardening, access control, surveillance of blind spots, ballistic protection, and mitigation of advanced threats like 3D-printed weapons.

Simulates a Linux terminal environment. Executes user commands (including natural language instructions) and returns raw output strictly within a single code block without explanations or conversational filler.

Evaluates a single AI response against a prompt using a 0-100% scale, grounded in rigorous criteria of Harmlessness, Honesty, and Helpfulness.

Implements a hierarchical document analysis system combining RAPTOR's recursive tree structure with semantic chunking and MixedBread embeddings for deep offline analysis of unstructured data.

一个结构化、分阶段的心理危机干预宏观协议，聚焦安全稳定、资源激活与成长巩固三个递进阶段，适用于高危但非急性自杀状态的青少年来访者。

一种制度化、结构化的跨系统协作治理框架，用于在学生存在自伤自杀高风险且家庭功能严重受损时，整合学校、家庭、社区、医疗和警方资源，同步动态信息、校准干预策略、明确责任分工并闭环追踪行动项，以守住生命安全底线。

一种融合精神分析客体关系理论与虚拟现实技术的跨流派干预框架，通过VR模拟父亲、母亲及重要他人等原型形象，引导社交焦虑来访者在虚拟场景中直面核心恐惧、唤起早年关系体验，并尝试修复或创造新的互动体验，从而超越行为训练层面，触及焦虑的深层心理动力根源。

基于玛格丽特·马勒分离个体化理论，对12–18岁青少年来访者在共生、分离与个体化阶段的发展任务完成情况进行结构化评估，识别其人际关系冲突背后的内在发展性动力；输出三阶段状态判断及对应行为证据锚点。

一个结构化的心理动力学取向咨询流程，划分为评估与概念化、关系建立与情绪宣泄、模式觉察与功能示范、内化与演练四个递进阶段，用于指导跨会话的长期咨询节奏。

一个结构化的心理动力学长程治疗流程，划分为关系建立、解释与经验重构、修通与扩通、结束四个递进阶段，每个阶段有明确目标、技术组合与时间分配。

在短程动力学治疗结束阶段，通过结构化、去病理化的语言明确告知来访者‘未来有需要时可随时重返咨询’，以此转化分离焦虑、预防被抛弃感，并支持治疗后持续修通。该协议非一般性告别，而是整合动力学分离议题处理、依恋修复与自我调节支持的会话干预。

在咨询中协同求助者设计并实施现场人际反馈实验（如邀请同伴当面评价其社交表现），用真实外部信息直接挑战其核心不合理信念。