OpenCode Plugins (2026): Which Still Work on OpenCode 2

The best OpenCode alternatives are the terminal coding agents developers say they moved to in blog posts and Hacker News threads, checked against independent hands-on reviews. If OpenCode feels bloated, get Pi, or oh-my-pi if you want Pi with the tools already built in. If you want one vendor's best models at a flat monthly price, get Claude Code or Codex CLI. Codex is the one that sandboxes commands by default. If you want open source with any model and no single company in charge, get Goose. If you want the nicest terminal interface, get Crush. If you already pay for GitHub Copilot, its CLI is included. A few things the makers' pages won't tell you. OpenCode 2 has a new plugin API, so OpenCode 1 plugins don't run on it, and that is a reason recent switchers give on Hacker News. Crush grew out of an older Go project that was also called OpenCode, and its license is source-available, not open source. Gemini CLI stopped serving personal accounts on June 18, 2026, and Antigravity CLI replaced it. Aider has had no commits since May 22, 2026. Kilo CLI is built on OpenCode, so switching to it changes little. A free agent still needs a model: you pay a provider per token or run one on your own hardware. The list is ordered by how many independent developers and reviewers recommend each agent, then by GitHub stars, then by price. Prices checked October 11, 2026.

The best extensions for the Pi coding agent (some people call them plugins), picked from the setups developers publish and the write-ups that tested them, not from download counts alone. Pi is deliberately bare: no sub-agents, no plan mode, no permission prompts. You add what you miss with one command, pi install npm:<package>. If you are starting out, three cover most of the gap with Claude Code: pi-subagents for delegation, pi-web-access for search, and rpiv-ask-user-question so the model asks before it guesses. Add Plannotator if you want plan mode, and pi-permission-system if you don't want the agent free to run anything. Two things have changed that older guides miss. Pi has had MCP support built in since version 0.99, so pi-mcp-adapter is now an upgrade and no longer a requirement. And several popular names have lookalikes: npm has at least four other packages named pi-mcp-adapter under other people's scopes, and more than one /btw extension, so install by the exact name shown here. Extensions run with the same access to your machine that Pi has. Pi's catalog lists more than 5,000 packages, so read the source of anything that touches your shell or credentials, and pin versions you depend on. Ordered by how many independent users and reviewers recommend each one (Kay Rhodes's setup guide, Composio's tested top ten, developers on Hacker News and Pi's own GitHub discussions), then by npm downloads and GitHub stars. Downloads count automated installs too, so treat them as a rough signal. Numbers checked October 10, 2026.

The best Claude Code plugins, picked from the setups developers publish after pruning their own installs and from Hacker News threads where people say what they kept, not from install counts alone. A plugin is a bundle. One install can carry skills, slash commands, subagents, hooks and MCP servers. A skill is a single instruction file Claude loads when it judges it relevant, and an MCP server is a connection to an outside tool, so a plugin can contain both. Type /plugin inside Claude Code to browse Anthropic's official marketplace, which is connected from the start, or run /plugin install name@claude-plugins-official. For anything else, add the maker's marketplace first with /plugin marketplace add owner/repo. If you install one thing, make it Superpowers, plus the LSP plugin for your language. Add Security Guidance if you ship code you don't read line by line, Plannotator if you want to mark up plans in a browser, and Claude HUD to see how full the context is. Every enabled plugin adds to the context on every turn, and several of the writers quoted here deleted more than they kept. Plugins run code on your machine with your permissions, and community marketplaces are not reviewed, so read what a plugin will install before you say yes. Two names trip people up. Claude Code has a built-in review command alongside the Code Review plugin. And Anthropic keeps a second, smaller marketplace in the anthropics/claude-code repo, where the loop plugin still goes by ralph-wiggum and not ralph-loop. The list is ordered by how many independent people recommend each plugin, then by installs and GitHub stars. Prices checked October 11, 2026.

The best Claude skills are the SKILL.md packages that Hacker News users and independent reviewers keep recommending. This list ranks collections and single skills you can install today, with what each adds and its catch. If you want Claude to plan and test before it codes, get Superpowers. If you want it to question you before it starts, get grill-me from Matt Pocock's skills. For better-looking pages, start with frontend-design and add Impeccable to review the result. To build your own skills, or to make real Word, Excel, PowerPoint and PDF files, use Anthropic's repository. For marketing work, use Corey Haines's pack. Skills install in three ways. In Claude Code, run /plugin install from a marketplace, or copy a folder into ~/.claude/skills. For most coding agents, npx skills add owner/repo does the copying. In the Claude apps, upload a ZIP under Customize > Skills. Skills follow the open Agent Skills format, so most of these also run in Codex, Cursor and Gemini CLI. Older guides are now wrong in places. Claude Code has merged custom commands into skills. Superpowers and Matt Pocock's skills are in Anthropic's official marketplace, so adding a separate marketplace is no longer needed. The Karpathy repository moved to a new owner. Star counts are huge and say little, and a link list such as awesome-claude-skills installs nothing. A skill is instructions plus scripts that run with your agent's permissions. Snyk scanned 3,984 skills from ClawHub and skills.sh and found a critical security issue in 13.4% of them. Read a skill before you install it, and install a few, not all. Items are ordered by how many independent people recommend each one, then by GitHub stars. Everything here is free. Numbers checked October 11, 2026.

The best OpenClaw skills, chosen from what developers and reviewers who run the agent say they keep installed, then checked against each skill's source and its ClawHub security audit. For email and calendar, start with Gog. For code, GitHub. For long articles, videos and podcasts, Summarize. For notes, Obsidian or Notion. To stop the agent repeating a mistake, self-improving agent. For cleaner prose, Humanizer. Safety first. ClawHub, the registry, lets anyone with a GitHub account publish, and a skill acts with whatever access your agent has. Security firms have counted more than a thousand malicious skills there since January 2026, and in June Unit 42 and Trail of Bits each showed skills getting past its scanners. OpenClaw's own June data labelled 25,504 of 67,453 skill versions suspicious and 206 malicious. Every skill page now has a security audit with an outcome such as Pass or Review. Read it, and the SKILL.md, before you install. Older guides are out of date in three ways. Most of the skills they tell you to install now ship with OpenClaw and switch on when their command-line tool is present. Web search, self-learning and memory search are built in, and Tavily has a first-party plugin, so the Tavily and QMD skills are no longer needed. And names are copied: at least five publishers use the slug gog, so install with the owner, as in @steipete/gog. Skill Vetter and Capability Evolver are on most lists and left out here. Skill Vetter's publisher links to a GitHub profile that no longer exists. ClawHub's audit marks Capability Evolver for review because it can run jobs handed to it by a remote server. Skills are ordered by how many independent reviewers and users recommend them, then by downloads or stars. All are free. Numbers checked October 11, 2026.

The best local LLM for coding depends on how much memory your machine has. This list ranks open-weight models you can run yourself, by independent benchmark runs and by developers who use them in coding agents. With 24 to 32 GB, start with Qwen3.8-27B. Its 17 GB 4-bit file matched the full model on Terminal-Bench 2.1 in Quesma's test. With 128 GB of unified memory, Qwen3.8-Flash-Next posted the top score in WonderRico's benchmark. With 16 GB, gpt-oss-20b fits whole on the card, and one developer reports useful work from a 2-bit Qwen3.8-27B. With 8 GB, try Qwen3.5-9B and keep the jobs small. At 64 GB, run the 27B models at 8-bit, a 29 GB file, with a longer context. The step above that needs about 100 GB. A coding agent only works if the model emits tool calls your server can parse. Gemma 4 shipped with broken tool calling that was fixed later, and Qwen chat templates have needed community fixes, so update Ollama, LM Studio or llama.cpp before blaming the model. Agents also need room: LM Studio loads models with an 8,192-token context unless you raise it. Scores on a model card are the maker's own. A mixture-of-experts model computes with a few billion parameters per token but still has to hold all of them in memory, and context takes memory on top of the file. Qwen3.8 thinks at its highest effort by default, which is slow on home hardware. Qwen3.8-Flash-Next is not Apache licensed. Older guides still recommend Qwen2.5-Coder, Qwen3-Coder and Devstral, which general-purpose models beat in the hands-on tests cited here. The list is ordered by how many independent testers and developers recommend each model for coding, then by downloads. Numbers checked October 11, 2026.