AI Agent2025
Compiles an agent's raw chat logs, documents, and tool traces into three persistent layers — index, learned skills, and user memory — so context survives sessions. Claims 92% Locomo-benchmark accuracy and up to 95% lower token cost than replaying history.
AI Agent2025
Collaborates on web tasks in real time: edit its plan before it runs, pause and grab the browser mid-task, and approve irreversible clicks before they happen. A research prototype for studying human-in-the-loop oversight instead of full autonomy.
MCP Server2025
Provides a long‑lived, in‑process file and content search library for editors and AI agents, with typo‑resistant fuzzy matching, frecency‑ranked results, background watchers, and a lightweight in‑memory content index — optimized for repeated searches in long‑running processes.
MCP Client2025
Provides ultra-fast, typo-tolerant file search and grep tuned for Neovim and AI agents, with built-in memory (frecency, git status, size, definition matches). It reduces agent token use and speeds developer file discovery in large repos.
AI Client2025
Indexes any repo into a knowledge graph of dependencies, call chains, and execution flows, then feeds it to AI coding agents via MCP so they stop missing context. Ships as a CLI plus a zero-install browser graph explorer with chat.
AI Image2025
A 20B-parameter MMDiT diffusion model that generates and edits images with accurate embedded text, including dense Chinese and English typography. Handles complex multi-line layouts and identity-preserving edits while keeping text legible.
MCP Server2025
An MCP (Model Context Protocol) server that lets AI assistants interact with Xiaohongshu (RedNote): check login, publish image/text or video posts, search and fetch feed/details, and manage comments — exposes HTTP+MCP endpoints and integrates with MCP clients via local Docker or browser automation.
AI Coding2025
Reviews each pull request for security issues: Claude reads the diff and flags vulnerabilities like injection, auth flaws, and hardcoded secrets as inline comments, with built-in false-positive filtering. Ships as a GitHub Action or slash command.
MCP Server2025
Helps developers connect IDEs and AI agents to the Figma MCP server to extract design context, generate code from frames, and write updates back to Figma. Includes client setup (VS Code, Cursor, Claude Code, Gemini CLI), tools and skills, best practices, and beta rate-limit notes.
AI Coding2025
Adds a lightweight, spec-driven workflow so AI coding assistants agree on requirements before code is produced — creates per-change artifacts (proposal, specs, design, tasks), exposes CLI slash-commands, and integrates with 20+ tools for repeatable AI-driven development.
AI Agent2025
Deploys autonomous AI agents that dynamically attack running apps and return validated proof-of-concept exploits instead of static-analysis noise. Specialized agents cover IDOR, injection, SSRF, XSS, and auth flaws, with HTTP proxy and CI/CD hooks.
Seven-week course that builds a production RAG system from scratch — an arXiv paper assistant that starts with BM25 keyword search, then layers hybrid vector retrieval, local-LLM generation, Langfuse monitoring, and an agentic LangGraph Telegram bot.