AI & LLM Integration for Claude Code

Capture a writer's voice from a few sample articles and rewrite text to sound like that author. Works for any author, 王建硕 included.

Time-aligned TTS dub of a video against a target-language SRT. Volcano TTS for Chinese, edge-tts for everything else. Single-speaker by default; opt-in visual diarization for multi-speaker.

wjs-evaling-voicedrop-prompts

Audio/video → timestamped SRT in the source language. Chinese uses Volcano ASR; other languages use OpenAI Whisper with word-level timestamps + punctuation-bounded re-segmentation.

VoiceDrop 配套 skill。两个模式 — (1) list：列出你在 VoiceDrop R2 上的全部文件（录音 / 文章 JSON / 标记）；(2) distill：从你的样本文章蒸馏文风，生成可贴入 VoiceDrop「设置→文风」的精简规则集，并可直接通过 Files API 上传。触发词："voicedrop list"、"列出 VoiceDrop 文件"、"voicedrop distill"、"蒸馏 VoiceDrop 文风"、"/wjs-voicedrop"。

Comprehensive red teaming methodology for both cybersecurity (MITRE ATT&CK, pentesting) and AI/LLM systems (prompt injection, jailbreaking, OWASP Top 10 LLM). Includes threat modeling, attack simulation, compliance validation (NIST, EU AI Act), and progressive attack technique references.

Agent capability frameworks: STAR framework for understanding implicit intent (framework-initiative), critical thinking architecture with reasoning selection (framework-critical-thinking), and adversarial review protocol with mandatory bug quota (adversarial-review). Build smarter, self-correcting agents and validate plans before execution.

Enables texting Claude directly from your iPhone via iMessage for on-the-go conversational AI access.

Help coding agents integrate with the Shuriken trading platform via its API, SDK, and CLI.

WakaTime plugin for Claude Code - Track your AI usage, lines of code generated, and time spent prompting AI

Audio processing skill for transcription, summarization, and content extraction from voice recordings and audio files.

SwiftUI preview integration for AI coding. Capture, review, and iterate on SwiftUI views using axe CLI.

Enhanced web browsing for Claude Code. Handles search, scraping, login flows, and dynamic pages via Playwright. Use for any task requiring real browser interaction.

Gives Claude Code a multi-agent coordination layer over MCP — an atomic shared blackboard (locked propose/validate/commit), permission gating with HMAC/Ed25519-signed tokens, federated token-budget tracking, and append-only audit-log queries. Backed by an orchestrator with 32 framework adapters (LangChain, AutoGen, CrewAI, Gemini, and more).

Real-time communication and voice agent integration with LiveKit. Enables AI agents to participate in video, audio, and messaging sessions.

Self-learning AI agents that improve over time. Features agent memory (tracks last 100 executions), pattern recognition, community standards contribution, and workflow optimization. Includes security, quality, deployment, infrastructure, and compliance workflows.

AI Operations Team — 9 agents: Deploy, Eval, Trace, Guard, Budget, Token, Prompt, Embed, Rank

AI Operations Team — Budget: LLM spend tracking, model cost optimization, budget alerts, and token efficiency audits across AI workloads.

Design cost reduction strategies — model tiering, prompt compression, caching, batch inference.

Contribute a discovery back to the upstream tonone repo — new skill, improved agent prompt, better routing rule. Use when you hit a gap, found a better pattern, or built something reusable that others would benefit from.

ML/AI engineer — model training, MLOps, feature engineering, LLM integration

Design and implement an AI feature integration — model selection, architecture pattern, system prompt, data flow, error handling, cost estimate. Use when asked to "add AI to this", "LLM integration", "add Claude/GPT", or "AI-powered feature".

Build an ML pipeline — from data to trained model to serving endpoint. Use when asked to "build ML model", "train a model", "prediction pipeline", "classification", or "regression".

Build a production-ready prompt package — system prompt, few-shot examples, output format, edge case handling, eval criteria. Use when asked to "prompt engineering", "build a prompt", "write a system prompt", or "improve this prompt".

ML reconnaissance — inventory all models, pipelines, data sources, and monitoring. Use when asked "what ML do we have", "model inventory", or "ML assessment".

AI Operations Team — Deploy: Model serving, inference APIs, blue/green deploys, rollback, and canary releases for production ML.

Design and configure model serving infrastructure — endpoint scaling, batching, GPU allocation.

Edge skill: edge-recon

Edge skill: edge-route

AI Operations Team — Embed: Embedding model selection, vector pipeline design, similarity search, and production index management.

Design embedding pipelines — model selection, batching, normalization, index refresh strategy.

Optimize similarity search — ANN index tuning, hybrid search, reranking, query expansion.

Design eval harnesses — task schemas, metrics, dataset versioning, eval-as-code patterns.

AI Operations Team — Evals: Eval harness design, benchmark suites, automated regression, human eval orchestration.

AI Operations Team — Guard: Input/output safety filters, PII detection, content moderation, and AI policy enforcement at runtime.

Design guardrail layers — input classifiers, output validators, PII scrubbers, policy rule engines.