发现 Skills

Find and identify the CBaseEntity_TakeDamageOld function in CS2 binary using IDA Pro MCP. Use this skill when reverse engineering CS2 server.dll or libserver.so to locate the TakeDamageOld function by searching for the TakeDamageOld debug string references and analyzing cross-references.

Find and patch the water jump velocity inside CCSPlayer_MovementServices_CheckJumpButton in the CS2 binary using IDA Pro MCP. This patch changes the water jump velocity from 100.0f to 145.0f by modifying the immediate operand of a mov instruction. Use this skill when reverse engineering CS2 server.dll or libserver.so to locate and patch the water jump height value. Trigger: water jump patch, CheckJumpButton water velocity, jump height patch, water jump 145

Generate this week's main text: planning (including Bloom taxonomy/review bridge/superline/roles) → scene-driven drafting → deep polishing → four-dimensional QA → revision loop → finalize and produce QA_REPORT.

Generates a complete, customized Non-Disclosure Agreement with plain English annotations, tailored to the specific parties and situation

Optimize website and web application performance including loading speed, Core Web Vitals, bundle size, caching strategies, and runtime performance.

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/PRD asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

Ensures new designs align with existing project architecture patterns to prevent drift and technical debt. Use when proposing or reviewing architecture.

Build Mojo packages (.mojopkg files) for distribution. Use when creating distributable libraries or during packaging phase.

Expert-level Vercel AI SDK v5 patterns for production chatbots. Use for: (1) Chat persistence with Drizzle/PostgreSQL, (2) Generative UI with typed tool parts, (3) Human-in-the-loop tool confirmations, (4) Custom data streaming with reconciliation, (5) Anthropic provider with extended thinking/reasoning, (6) Type-safe message metadata with token tracking. Covers advanced patterns only - assumes basic AI SDK knowledge. NOT for AI SDK v6.

Create new AiderDesk UI themes by defining SCSS color variables, registering theme types, and adding i18n display names. Use when adding a theme, creating a color scheme, customizing appearance, or implementing dark mode and light mode variants.

Comprehensive performance optimization agent that provides expert guidance on profiling, frontend optimization (Core Web Vitals), backend optimization, caching strategies, and performance monitoring. Covers browser rendering, network optimization, database tuning, memory management, and load testing. Use for performance audits, optimization initiatives, or establishing performance budgets.

ASCII video: convert video and audio into colored ASCII MP4/GIF.

Hugging Face hf CLI: search, download, and upload models and datasets.

Pitfalls, examples, and edge cases for Hermes Kanban workers. The lifecycle itself is auto-injected into every worker's system prompt as KANBAN_GUIDANCE (from agent/prompt_builder.py); this skill is what you load when you want deeper detail on specific scenarios.

Outlines: structured JSON/regex/Pydantic LLM generation.

Consumer product shopping research and evaluation workflow — from market research to three-end delivery (Feishu + NAS file system + NAS Obsidian). Trigger conditions: user says "help me do a shopping research and evaluation" or "help me research XX product". Includes: market landscape research, official site specification scraping, YouTube/Bilibili video subtitle extraction, PDF report generation, and dual-end delivery.

Songwriting craft and Suno AI music prompts.

Execute plans via delegate_task subagents (2-stage review).

TRL: SFT, DPO, PPO, GRPO, reward modeling for LLM RLHF.

vLLM: high-throughput LLM serving, OpenAI API, quantization.

Multi-phase test suite optimization with adversarial review

Use this skill when the user needs to integrate with the Hubble Network Platform API for IoT device management, packet data retrieval, webhook configuration, metrics tracking, user management, or billing. Triggers include 'Hubble API', 'Hubble devices', 'Hubble packets', 'register IoT devices', 'configure webhooks for Hubble', 'Hubble Network integration', and 'stream packet data'.

深度客户开发工具集。当需要：发现企业邮箱/联系人、验证邮箱有效性、深度爬取公司网站（含 JS 渲染）、搜索潜在客户公司列表、查询贸易统计数据（OEC/WTO/UN Comtrade）时使用。依赖本地 super-crawler 工具（~/projects/super-crawler），适合 B2B 销售、外贸开发、竞争对手调研等深度场景。日常搜索请优先使用 web_crawl 或 tavily_search；本技能用于需要获取联系人和贸易情报的深度客户开发场景。

Publish and manage research papers on Hugging Face Hub. Supports creating paper pages, linking papers to models/datasets, claiming authorship, and generating professional markdown-based research articles.