AI Engineer is presenting AIE CODE as its final flagship event of 2026, with two partner conferences also planned. The invite-only, single-track summit is intended for 1,000 agentic engineers and will recap leading coding themes from 2026 while previewing 2027. Its stated themes include “Scaling without Slop” and “System One models.”
36Kr, citing Crunchbase data, reported that global startup financing reached about $510 billion in the first half of 2026, exceeding the previous full year, with roughly 80% of first-quarter venture investment going to AI companies. The report also said OpenAI raised $122 billion in a single round at an $852 billion valuation, while Anthropic reached a $965 billion valuation. In an interview ahead of Hong Kong FinTech Week × StartmeupHK 2026, Huaying Capital founding managing partner Ji Wei and InvestHK Acting Head of Startups Lü Jinyong said AI competition is shifting from model parameters and compute build-out toward data flywheels, real customer demand, willingness to pay, sustainable cost advantages and cash-flow discipline for applications.
Expo released Expo Skills, a collection of official AI agent skills covering Expo projects, EAS services, React Native, and iOS and Android constraints. The skills can be installed through official plugin marketplaces for Claude Code and Codex, or through the skills CLI for Cursor, OpenCode, GitHub Copilot, Windsurf, Gemini, and other coding agents. The framework skills are open source under the MIT License, while Expo describes the collection as guidance for building, deploying, upgrading, and debugging Expo apps.
Cursor released its Agent SDK on October 20 for external use. The framework was used to build Cursor’s hosted Bugbot, Security Reviewer, and Rollouts agents, according to the report. Cursor said useful agents should be tailored to a team’s working style, while teams previously had to accept predefined workflows or build from scratch.
Developer EthanWng97 released pixtuoid v0.21 with a revamped 2.5D cutaway office that renders coding agents as pixel coworkers in terminals supporting kitty, iTerm2, or SIXEL images. Unsupported terminals receive the classic view, and the application can also run in a floating, always-on-top desktop window; it is available under the MIT License through Homebrew or npm.
HKUDS presented AutoSwarm, an installable system that coordinates existing AI agents as one team by selecting agents according to their capabilities, sharing context and memory, and connecting complementary tools and workflows. It can run subtasks in parallel and synthesize their results, with support across Windows, macOS, and Linux through desktop apps, CLI tools, IDE extensions, and skill-enabled environments. The repository also includes setup instructions, provider probes, and an accompanying technical paper.
The open-source AI Design Skills repository provides self-contained Markdown design instructions that can be used with Claude Code, Cursor, Codex, Windsurf, Cline, and other tools that read rules files. Its first listed skill, landing-page-design, covers page structure, conversion copy, typography, spacing, radii, and motion; users can install individual folders or copy the SKILL.md file into their preferred configuration.
Glasser, a data service from Super Intent, lets AI agents purchase premium data one API call at a time through a single key. The company says the catalog includes more than 2,000 paid endpoints from over 80 providers, with agents able to choose a source, check its cost, run the call, and report the charge and raw result link. TestingCatalog tested the service with Claude and found that a Hunter company lookup took under two seconds, while the article notes that its report was sponsored and several capability claims come from Glasser.
Huashu Design 3.0 is an open-source agent skill that turns prompts into web designs, clickable app prototypes, editable PPTX presentations, animations and infographics. Its README says it can generate six visual candidates for users to narrow down to three, provides an 84-style library with samples, and supports Claude Code, Codex, Cursor, OpenClaw and Hermes. The README also reports typical production times of 3 to 30 minutes and describes the output as resembling the work of a professional design team.
In an authorized TMTPost analysis, Zhixun Think Tank argues that model quality can win trials but does not necessarily create lasting user loyalty. It recommends comparing candidates on real task delivery, rework time, full cost, workflow integration, and continuity, rather than relying only on leaderboard positions.
The MIT-licensed higgsfield-ai/skills repository documents agent skills for image, video, 3D and audio generation, visual brand systems, product imagery, full-stack website development and deployment, narrated explainer videos, YouTube thumbnails, and Virality Predictor scoring for finished videos. It says the skills work with Claude Code, Cursor, Codex, and other agents that load Markdown-based skills. Installation options include npx, GitHub CLI, the Claude Code marketplace, and a setup script. These capabilities and compatibility claims are vendor descriptions and were not independently tested in the supplied report.
The obsidian-second-brain project announced Crew, a shared logbook that lets Claude Code, Codex, Cursor and Grok Bot review completed jobs and leave lessons for one another. The project says Crew can store records in any plain folder, operates without Obsidian and requires no background loop; lessons become rules only when a human runs the keep command.
DeepSeek V4.1 Flash is now available in Cursor and can be enabled through Cursor Settings > Models. The report says every effort level sits on the CursorBench cost frontier, with costs ranging from $0.60 to $0.98 per task, and that no model scores higher.
The author of LLM Tails says the MIT-licensed project combines Ponytail and LLM Wiki through prompts, hooks, and skills that encourage coding agents to read first, write minimally, and verify claims. The author says it has no dependencies and works with Claude Code, Cursor, GitHub Copilot, Codex, Gemini, Windsurf, Cline, and Kiro. Claims about smaller codebases, lower token use, and fewer hallucinations are not supported by numerical results, while the referenced same-prompt comparison image is not available in the supplied report.
Entire is now available to developers as a GitHub alternative designed around coding agents and a decentralized Git network. Its Trails workflow replaces traditional pull requests with agent-scored evaluations, while the open-source Entire CLI, Continuous Memory, and multi-location private repositories support agent-assisted development. Entire claims Git hosting can be up to 89 times faster than GitHub and that Continuous Memory can reduce token use by as much as 71%, but these performance figures have not been independently verified.
Author kyriewen says they screened more than 1,400 Claude Code Skills and retained five used in their daily workflow: Superpowers, Taste Skill, Graphify, Deep Research and find-skills. The article argues that useful Skills alter how an AI coding tool plans, debugs, designs interfaces, researches topics or discovers other tools, rather than merely repackage a prompt; its performance claims and GitHub popularity figures are not independently verified here.
Experiential Labs describes Experiential as an open-source gateway and router for agent workflows. Its README says the gateway provides one OpenAI-compatible API for hosted, bring-your-own-key and local models, with controls over model access and spending by user, agent and use case. The project also says users can turn production traffic into custom routers or specialized models, and offers hosted and local gateway options. These capabilities are described by the project and were not independently verified.
A 红杉汇 article republished by 36Kr argues that AI agents are shifting software value away from basic feature sets toward scarce data, hard-to-replace workflows, developer distribution, security, identity and compliance capabilities. It cites reported 2026 transactions involving Cursor, Hugging Face, OpenRouter and World Labs, alongside Airtable's reported sale to Bending Spoons for about $1.3 billion, compared with Airtable's peak valuation above $11 billion in late 2021. The article presents these cases as evidence of changing acquisition-market priorities.
PaperPilot v0.6 is an Apache 2.0-licensed PDF and document toolkit written in Rust. Its shared engine is available through MCP, a CLI, WASM, a Tauri desktop app, and a local REST API. Current capabilities include PDF organization and optimization, text extraction, OCR through PaddleOCR, watermarking, signature validation and removal, PII scanning, deep sanitization, and Excel-to-JSON or CSV conversion. The project says files are processed locally or in browser memory and are not uploaded. Its roadmap lists features for later versions, including format conversion, encryption, annotations, rendering, repair, comparison, PDF/A support, and Bates numbering; these are not part of v0.6.
Hallmark v2 is an MIT-licensed design skill for Claude Code, Codex CLI, Cursor and OpenCode that aims to reduce stereotyped AI-generated visuals. According to Nutlope’s repository, the update adds 48 themes, custom design requests, URL and development-server screenshots, redesign and audit commands, and browser-based checks for machine-like design patterns. Browser checks require Puppeteer Core and Chrome, while static checks remain available without them.
In a Juejin commentary, the author argues that GPT-5.6, K3 and Claude Fable5 can handle many basic development steps that older Skills packages explicitly encoded. Using Codex as the main example, the author recommends retaining only Skills that add scripts, templates, specialist knowledge, personal preferences or reusable checks, while placing routine project rules in AGENTS.md and mechanical safeguards in Hooks, CI, linters or tests.
Open Duet is an open-source, self-hosted workspace for creating named AI agents with instructions, roles, workspaces, memory, routines and approval settings. Its README says agents can delegate tasks, retain context and run scheduled work. Users can connect engines such as Claude Code, Codex CLI, Gemini CLI, Cursor, OpenRouter, Ollama or LM Studio. The project does not include model access or subscriptions. Team delegation requires Claude, while the self-hosted server, Docker-based agent computers and browser tools remain experimental; the server setup is intended for a single owner rather than a multi-user organization.
SPACEXSI reported on X, citing Cursor, that the coding tool now has Computer Use capabilities for operating other applications on a user's laptop. The report said the feature can also run mobile simulators for app development and can be activated on demand or by @-mentioning the “Computer Use” command. The report did not specify rollout timing, supported operating systems, availability restrictions, permissions, or safety safeguards.
A Juejin post by developer "程序员鱼皮" reviews REA (Reverse Engineer Anything), an open-source command-line tool plus MCP service that wraps established reverse-engineering utilities such as Ghidra, IDA and JADX so AI coding assistants like Cursor, Codex and Claude Code can drive them. The author reports that Codex with GPT-6 Astra successfully added a new character to an old Android game, while Claude Opus 5.5 refused to fully decompile a Flutter app on safety grounds and a switch to DeepSeek produced a working skeleton analysis. He concludes REA invents nothing new, that its roughly 170 tool definitions and 100k-plus tokens of context are heavy, and that models can reverse-engineer without it; the post also repeats anecdotal, unverified user stories and a claim that the project's GitHub stars jumped past 200,000 in early October.
Reports indicate that Microsoft is testing Project Pantheon under the name Context Vault, a Windows 11 AI information app that stores content users manually add rather than automatically taking regular screenshots. Users can save webpages, files, screenshots, text, audio and video, then use AI to organize, search, summarize and ask questions about the material. The reports describe Windows entry points, including the File Explorer sharing menu and Click to Do, as well as Model Context Protocol access for tools such as Claude and Cursor. One report says the app must be installed from the Microsoft Store, while another describes deeper Windows integration; the app is not officially listed and its release remains uncertain.
An InfoQ article authorized for publication by 36Kr reviews vulnerabilities affecting GitLab, Chrome, Node.js, PHP, LiteLLM, eMudhra and Railway, alongside separate incidents involving AI-assisted attacks on Korean financial institutions and an AI agent reported to have deleted PocketOS’s production database and backup in nine seconds. The report says the listed vulnerabilities may not be technically related, does not establish a coordinated campaign, and does not claim that AI discovered every flaw.
Latent Space describes Claude Mods as a plugin system that lets users customize Claude Code’s behavior and appearance through TypeScript or JavaScript functions. The author reports using it to adapt the coding agent to a personal engineering workflow and says Claude Code can help create mods through a built-in skill. The article also notes that mods run with the user’s permissions and are not sandboxed, so installing third-party code requires careful review.
SpaceX reported Q3 revenue of $14.58 billion, up 161% year over year, and net profit of $4.4 billion versus a $1.1 billion loss a year earlier, according to CNBC and The Next Web. Its Super Intelligence unit generated $7.56 billion in revenue, surpassing Starlink’s $5.25 billion, while AI operating income reached $431 million. SpaceX spent about $19.6 billion on capital expenditure during the quarter, including $16.4 billion for AI and data centers.
Asymptote Labs raised a $5.5 million seed round led by Tower Research Ventures, with Lytical Ventures, 1984 Ventures and South Park Commons also participating. Its open-source Beacon tool records prompts, commands and file edits performed by more than 20 coding agents, while a paid platform can block risky actions such as data leaks, according to the company. The valuation was not disclosed, and the funds will support engineering, hiring and sales.
What problem it solves
Web teams and developers can need to rebuild an existing site in a modern codebase when migrating platforms, recovering lost source code, or learning how a production site is constructed. This template gives an AI coding agent a repeatable workflow for recreating a site as an editable Next.js application while comparing its appearance and behavior with the source.
Key featur…
What problem it solves
AI agents often produce plausible SQL without the business definitions, approved joins, units, and examples needed to make results trustworthy. Teams also need a reviewable way to manage that context and turn governed answers into shareable dashboards across existing data sources.
Key features
Governed semantic and context layers — Business meaning is represented through the…
What problem it solves
Personal and team knowledge is distributed across notes, conversations, trackers, and research, making it difficult to capture, organize, verify, and synthesize useful information. COG provides a markdown-based workflow in which agents can maintain a shared knowledge base, support recurring research and operational tasks, and verify selected work through independent checks.
…
What problem it solves
AI agents need product-specific instructions to use NVIDIA software reliably across physical AI, robotics, simulation, CUDA-X libraries, RAG, and platform workflows. The catalog centralizes these skills while keeping them maintained in their respective product repositories and synchronized through an automated pipeline.
Key features
Agent-discoverable skill catalog — The NVI…
What problem it solves
Developers and coding agents need to understand how code is connected before making changes. GitNexus provides a graph-based view of symbols, callers, dependencies and execution flows to support code discovery, impact analysis and refactoring.
Key features
Graph-based code intelligence — Index a codebase as a knowledge graph covering symbols, calls, dependencies and executio…
What problem it solves
Agent memory is often opaque: stored context cannot be easily inspected or understood. OpenViking addresses this by organizing knowledge, memory, and skills in a browsable virtual filesystem, with summaries that help agents identify relevant context before reading full files.
Key features
Unified context filesystem — Knowledge, memory, and skills share one filesystem address…
What problem it solves
Traditional development accumulates technical debt and leaves important knowledge scattered, forcing later changes to rediscover context. The plugin addresses this by turning each engineering cycle into a structured workflow and writing durable learnings into artifacts that subsequent cycles can use as grounding.
Key features
Structured engineering loop — A six-step cycle mo…
What problem it solves
The hidden system instructions that shape AI product behavior are difficult to inspect and change across models, products, and versions. This repository makes collected prompts available in a structured, product-oriented archive for study and comparison.
Key features
Broad product coverage — The archive covers prompts from major AI providers and products, including Anthropic…
What problem it solves
Chrome DevTools for agents (`chrome-devtools-mcp`) lets your coding agent (such as Antigravity, Claude, Cursor or Copilot) control and inspect a live Chrome browser. Both give your AI coding assistant access to the full power of Chrome DevTools for reliable automation, in-depth debugging, and performance analysis.
Key features
Browser control and inspection — Give coding age…
What problem it solves
AI agents need relevant design context from Figma files to generate code that follows a design system and reflects design intent. This project documents how to connect the Figma MCP server to supported clients, use its tools and skills, and apply practices for more consistent results.
Key features
Connect supported MCP clients — Setup instructions cover the Figma MCP server …
What problem it solves
Coding agents need a structured way to create, manage, and build from Google Stitch designs while integrating with the Stitch MCP server. The project organizes design, code generation, and supporting workflows into skills that can be installed with supported coding agents.
Key features
Design workflows — The design plugin covers converting frontend code into Stitch designs, …
What problem it solves
AI coding agents lose context between sessions, so the same decisions get re-explained and nothing about the project is remembered. They also lack a shared workflow and evidence, leaving the human to read every line to confirm what was done. This project adds memory, a workflow, and evidence to agents you already have.
Key features
Persistent project memory with Engram — Eng…
What problem it solves
Developers need a consistent way to control locally configured coding agents across mobile, web, and desktop environments. T3 Code addresses this by providing multiple client applications, support for several agent providers, and remote access to a local server.
Key features
Control from mobile, web, and desktop — Use the mobile, web, or Electron-based desktop application to…
What problem it solves
Teams that rely on AI agents need a way to give each agent a role, workspace, and instructions, keep context across conversations, and control what agents are allowed to do. Without that, agents run blind: actions happen without review, memory does not persist, and repeating scheduled work means redoing setup by hand.
Key features
Create agents and teams — Give an agent a na…
What problem it solves
Coding tools use different component formats, model names, permissions, and discovery rules. Contributors need a shared way to write reusable plugins, agents, skills, and commands once while producing compatible artifacts for each supported harness.
Key features
Multi-harness component support — One collection of plugins, agents, skills, and commands can be used across Claud…
What problem it solves
AI coding assistants can generate interfaces with visual and structural patterns that feel machine-made, such as repetitive headings, empty sections, misplaced heroes, and forms that do not complete their intended action. Hallmark applies design constraints, product-aware page structures, browser-based linting, and fresh-context review to help generated interfaces look desig…
What problem it solves
Developers need a coding agent that can work directly from their local development environment without requiring a separate desktop application. The project provides installation paths for terminals, IDEs, and desktop use, along with authentication through a ChatGPT plan or an API key.
Key features
Local terminal coding agent — Run Codex CLI locally on your computer to use a…
What problem it solves
Agent workflows, corrections, debugging patterns, and repository conventions often disappear when a session ends or remain trapped inside one coding tool. Beacon addresses this by capturing session history across harnesses and turning useful findings into reusable memory for future agents.
Key features
Cross-harness session history — Capture session history from Claude Code,…
What problem it solves
DeepSeek Harness workspaces, tools, and project setup normally remain tied to the computer running the Harness instance. The project addresses the need to continue using that environment from another computer, phone, or browser without requiring a public listening port or public IP on the Host.
Key features
Remote workspace access — Open local workspaces from a web browser, …
What problem it solves
Developer workflows often need consistent access to external tools, reusable agent instructions, and service-specific procedures. This repository packages those capabilities as standalone plugins with standardized manifests and supporting rules, skills, and MCP server definitions.
Key features
Developer workflow assistance — Plugins support tasks such as code review, reposit…