Discover AI Agent Skills

Bridges Claude Code and Codex sessions when native tools can't reach the target. Use for 跨产品通信 / 脚本回帖 / 送达排查: cross-product messages, hooks or scripts posting to a session, verifying delivery (Held messages included), shared-work ownership, or an inbound peer assertion, or for cross-machine pairing. Not for ordinary same-product or parent/subagent messaging (use native tools), spawning agents, or moving full history.

Use when the user asks for fix my look or a task matching the examples below. Change ANYTHING inside a video — background, scene, lighting, outfit, weather, mood — from a free-form prompt, while keeping the EXACT original facial identity, motion, speech, audio AND closest supported output ratio. Edits the first frame with gpt-image-2, then propagates that look across the clip with Kling reference-video using the original clip as the identity anchor. Triggers: "change anything in my video", "edit my video with a prompt", "change the background of this video", "change my outfit in this clip", "restyle this video without changing the person", "put me on a beach", "make this video at night", "/fix-my-look".

Article illustrations: type × style × palette consistency.

Vendor-neutral framework for scoring software health, estimating technical debt, assessing cloud readiness and open-source safety, and communicating quality to business stakeholders. Use when you need to quantify code health at the application or portfolio level rather than fix individual findings.

Use cmux Cloud machines through the CLI. Use when the user says cmux cloud, cloud VM, cloud machine, cmux vm, Base workspace, machines panel, run/exec on a cloud machine, cloud agent, vm snapshot/fork/restore, push/pull files to a VM, open a port or desktop on a machine, or notify from inside a VM.

Structured workflow for summarizing code changes after completing tasks. Creates clear, actionable summaries of what was changed, why, and what to verify.

In-memory caching in Golang using samber/hot — eviction algorithms (LRU, LFU, TinyLFU, W-TinyLFU, S3FIFO, ARC, TwoQueue, SIEVE, FIFO), TTL, cache loaders, sharding, stale-while-revalidate, missing key caching, and Prometheus metrics. Apply when using or adopting samber/hot, when the codebase imports github.com/samber/hot, or when the project repeatedly loads the same medium-to-low cardinality resources at high frequency and needs to reduce latency or backend pressure.

Query NHGRI-EBI GWAS Catalog for SNP-trait associations. Search variants by rs ID, disease/trait, gene, retrieve p-values and summary statistics, for genetic epidemiology and polygenic risk scores.

This skill should be used when the user asks to "add MCP server", "integrate MCP", "configure MCP in plugin", "use .mcp.json", "set up Model Context Protocol", "connect external service", mentions "${CLAUDE_PLUGIN_ROOT} with MCP", or discusses MCP server types (SSE, stdio, HTTP, WebSocket). Provides comprehensive guidance for integrating Model Context Protocol servers into Claude Code plugins for external tool and service integration.

Guide identification, measurement, and management of operational risk in trading and brokerage operations. Use when designing trade error detection and correction procedures, investigating trade breaks and reconciliation failures, classifying loss events under Basel taxonomy, developing key risk indicators (KRIs) and dashboards, responding to system outages or data feed failures or order routing errors, conducting root cause analysis after a trade error or operational incident, planning business continuity and disaster recovery for trading desks, preparing for FINRA or SEC operational risk examinations, or assessing technology risk in OMS and market data systems. Also covers fat-finger errors, error account P&L, and corrective action tracking.

Master prompt engineering for AI models: LLMs, image generators, video models. Techniques: chain-of-thought, few-shot, system prompts, negative prompts. Models: Claude, GPT-4, Gemini, FLUX, Veo, Stable Diffusion prompting. Use for: better AI outputs, consistent results, complex tasks, optimization. Triggers: prompt engineering, how to prompt, better prompts, prompt tips, prompting guide, llm prompting, image prompt, ai prompting, prompt optimization, prompt template, prompt structure, effective prompts, prompt techniques

Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.

Best practices for Matplotlib data visualization, plotting, and creating publication-quality figures in Python

Manages project knowledge using ByteRover context tree. Provides two operations: query (retrieve knowledge) and curate (store knowledge). Invoke when user requests information lookup, pattern discovery, or knowledge persistence. Developed by ByteRover Inc. (https://byterover.dev/)

First-principles assumption auditor. Classifies each hidden assumption (fact / convention / belief / interest-driven), ranks by fragility × impact, and rebuilds conclusions from verified premises. Bilingual: auto-detects Chinese or English.

科特船长 - 办公自动化脚本，Excel/Word/文件批量处理

Agent skill for consensus-coordinator - invoke with $agent-consensus-coordinator

Apply codebase error handling patterns: Zod validation at boundaries, typed errors, early returns, and retry/backoff. Use when implementing error handling or input validation.