Every plugin for Claude Code we can find on GitHub, read straight from each repo's .claude-plugin manifests. Install one with claude plugin marketplace add owner/repo and claude plugin install name@marketplace — each page has the exact lines. Looking for Codex? See Codex plugins.
See what Claude costs you and spend less of it. In chat: hand a long chat over to a new one, digest long documents, set up Projects, check tokens, sharpen prompts. In Claude Code: Haiku 5.5 subagents for reading and test runs; /cco audit of your own transcripts. Local, no telemetry.
Use when the user needs academic paper retrieval — searching scientific literature and patent/author/source metadata, finding paper chunks for RAG-style citations, or expanding original text around a known paper offset. Provides six Sciverse tools (search_metadata, semantic_search, list_catalog, list_paper_relations, read_content, get_resource) via the sciverse-mcp-server MCP server.
Reverse-engineer any motion reference (video, GIF, X/Dribbble link) into production animation code — frame-level ffmpeg dissection, measured timing/easing/stagger, and a faithful React/TypeScript framer-motion port
Arcium Network for encrypted computation via MPC — data stays private during processing, enabling dark pools, sealed-bid auctions, encrypted voting, confidential DeFi, and hidden game state on Solana
Transform text scripts into professional videos with AI voiceovers, stock footage, and Remotion rendering. Advanced features include direct filesystem access, environment management, and whitelisted command execution.
Real-time voice interaction for Claude Code. Speak naturally while Claude works - interrupt, redirect, or provide continuous feedback without stopping.
Make videos drawn with code, such as motion graphics, launch and explainer videos, animated charts, logo reveals and social clips: write one HTML page that draws each frame from its time, render it to MP4 with Helios, check the frames, add a soundtrack.
Guides disciplined AI-assisted engineering that avoids cognitive surrender and keeps humans accountable. Use for non-trivial implementation, architecture, security, or operational tasks.
Local-first resolver for Hugging Face models (GGUF, MLX, safetensors). The agent checks your own storage and any mounted drives before downloading anything.
The Workbench Architect — turns one vague sentence into a build-ready Workbench Spec, before any UI exists. Covers both a personal daily-return page (养宠 · 记账 · 增肌) and a system workbench with modules, page trees and a data model (CRM · ERP · AI Agent 控制台 · 数据看板 · 智慧工厂). Asserts a whole workbench instead of running an interview, classifies by structure rather than audience (eleven of them), and binds every number in the hook to a field and a writer — so the page still has something to say on day two instead of rendering empty. Outputs .workbench/spec.md for finesse-ui, Cursor, v0 or a human.