vercel-labs/open-agents
10 skills1.2k installs57.5k starsGitHub
Install
npx skills add https://github.com/vercel-labs/open-agentsSkills in this repo
1Web Animation DesignDesign and implement web animations that feel natural and purposeful. Use this skill proactively whenever the user asks questions about animations, motion, easing, timing, duration, springs, transitions, or animation performance. This includes questions about how to animate specific UI elements, which easing to use, animation best practices, or accessibility considerations for motion. Triggers on: easing, ease-out, ease-in, ease-in-out, cubic-bezier, bounce, spring physics, keyframes, transform, opacity, fade, slide, scale, hover effects, microinteractions, Framer Motion, React Spring, GSAP, CSS transitions, entrance/exit animations, page transitions, stagger, will-change, GPU acceleration, prefers-reduced-motion, modal/dropdown/tooltip/popover/drawer animations, gesture animations, drag interactions, button press feel, feels janky, make it smooth. A comprehensive guide for creating animations that feel right, based on Emil Kowalski's "Animations on the Web" course.312installs2Plan ModePlanning doctrine skill requiring exploration, clarification, design, review, and presentation before implementation. Emphasizes thinking broadly while implementing narrowly, treating every change as a system modification affecting data flow, state ownership, interfaces, lifecycle, failure modes, and observability. Prefers authoritative derivable designs over transient client-only state. Five steps: explore codebase and lifecycle end-to-end with parallel agents when scope is uncertain; clarify ambiguities with the user; design comparing direct vs system-coherent implementations with eight explicit design questions; review against original intent and full lifecycle; present concise executable plan to PLAN.md. Triggers on /plan, plan this, or design an approach.163installs3Frontend DesignA frontend design skill from Vercel Labs open-agents that helps solo builders craft UI layouts and visual designs for AI-powered applications. Reach for it during the build phase when you need to rapidly generate or iterate on frontend designs for agent interfaces without starting from scratch.145installs4Emil Design Engemil-design-eng is a design engineering skill that applies the UI polish philosophy of Emil Kowalski to interface code. It covers animation decision-making, easing selection, GPU-safe property choices, and component-level interaction patterns like button press feedback. A solo builder reaches for it when a functional UI needs to feel crafted rather than functional — when animations, transitions, and micro-interactions need to be correct, not just present.137installs5Agent Browseragent-browser is a Vercel Open Agents skill that wraps a CLI for programmatic browser control—ideal when solo builders need to navigate sites, interact with DOM elements, fill forms, reload flows, capture evidence, or scrape structured data without maintaining a separate Playwright script repo. The documented workflow is deliberately small: open a URL, take an interactive snapshot to obtain element refs, perform clicks and fills by ref, then re-snapshot after navigation or heavy DOM churn. Commands cover full navigation (including connect 9222 for CDP), aliases for goto and quit, and page analysis via snapshot -i. It complements agentic coding loops where the model verifies UI behavior after a build. Complexity sits at intermediate because you must manage browser lifecycle, ref invalidation, and allowed bash tooling boundaries.135installs6Code ReviewThe code-review skill reviews code changes and provides actionable feedback when users ask to review PRs, diffs, commits, or code changes. It determines review scope from input: default uncommitted changes via git diff and git diff --cached, specific commits via git show, branch comparisons via git diff branch...HEAD, or pull requests via gh pr view and gh pr diff. After gathering diffs it reads full modified files because isolated hunks lack context, traces trust boundaries for auth, storage, networking, rendering, and secrets, and checks convention files like CONVENTIONS.md or AGENTS.md. Review priorities are bugs first including logic errors and edge cases, then security with concrete exploit paths for injection, XSS, auth bypass, and secret exposure, structure fit against existing patterns, and performance only when obviously problematic. Before flagging issues the skill requires certainty, reviews only changed code, and uses an explore agent to verify patterns. Output stays direct about severity and states attacker-controlled inputs for security findings. Triggers include /review, review this PR, review my changes, and code review requests.79installs7WorkflowThe workflow skill creates durable, resumable workflows using Vercel Workflow DevKit that survive restarts, pause for external events, retry on failure, and coordinate multi-step operations over time. Agents must read bundled docs in node_modules/workflow/docs before coding because installed API shapes differ from outdated training data. Core directives are use workflow on orchestration functions and use step on cached retryable units with full Node.js access. The skill recommends putting logic in steps and keeping workflow functions as thin orchestrators to avoid sandbox limits on fetch, setTimeout, and Node modules. Error handling uses FatalError for permanent failures and RetryableError with retryAfter for transient ones. Serialization rules require plain data only, not functions or class instances. Streaming uses getWritable in step functions with lock release. Framework integrations cover Next.js withWorkflow, Vite, Astro, and Nitro modules. Debugging commands include npx workflow health, npx workflow web dashboard, and npx workflow inspect runs. Triggers include workflow, durable functions, resumable, or workflow devkit mentions.58installs8Ai Sdkai-sdk is a reference skill from Vercel Labs that teaches agents how to call models through Vercel AI Gateway using the AI SDK—the fastest path for solo builders who want one HTTP surface instead of juggling per-vendor SDK keys. It covers authentication patterns (OIDC on Vercel, API keys locally), the two import styles for the gateway provider, and the operational rule to pull the current model catalog via curl before coding. That last constraint matters because model slugs change often and hallucinated IDs break builds quietly. Use it when you scaffold chat features, summarization endpoints, or agent loops in Next.js or Node backends destined for Vercel. It is intentionally narrow: gateway transport and SDK calls, not prompt design or evaluation harnesses.52installs9Deploy Open HarnessThe deploy-open-harness skill walks maintainers through deploying their own copy of Open Harness on Vercel with credential scoping tied to the current repository rather than legacy Harness assumptions. It requires reading README.md, apps/web/.env.example, and key auth, crypto, GitHub App, Redis, and sandbox source files before advising, trusting code over docs when they disagree. Users choose minimal deploy for Vercel sign-in and hosted UI only, or full deploy adding GitHub account linking, App installation, private repo access, pushes, and PR creation. Required credentials include POSTGRES_URL and JWE_SECRET for baseline runtime, plus ENCRYPTION_KEY and Vercel OAuth values for usable hosting, with GitHub App variables needed only for repo-backed flows. Safety rules forbid pasting secrets into chat and separate blockers for minimal versus full paths. Deployment flow covers fork, Vercel import, baseline env vars, first deploy for stable URL, OAuth app creation, redeploy, and optional Redis. First-run verification checks Vercel sign-in, session creation, and for full deploy GitHub linking and sandbox startup. Triggers include deploying, self-hosting, or configuring credentials for a.49installs10Remove Demo LimitsThe remove-demo-limits skill guides maintainers through stripping Open Harness hosted demo restrictions from a fork with the smallest coherent code change. It targets managed-template trial session caps, message caps, hosted-only restriction errors, demo delete restrictions, and any auth callback blocking non-Vercel emails on hosted deployments. Required inspection covers apps/web/lib/managed-template-trial.ts, session and chat API routes, message deletion route, and their tests, plus searches for identifiers like isManagedTemplateTrialUser, MANAGED_TEMPLATE_TRIAL_, deploy-your-own, and hosted restriction strings. Changes delete or simplify guards blocking session creation, chat messages, or deletion for managed-template users, remove dead helper code, and update tests asserting hosted caps without adding feature flags or new dependencies. Verification runs bun run ci after edits, clearly separating pre-existing failures from restriction removals. Output reports which restrictions were removed, files changed, and verification results. Triggers include remove demo limits, remove trial limits, remove hosted restrictions, or open this up for my fork. Use whenever fork maintainers nee.49installs