[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"extension-marketplace-vercel-labs-agent-browser-en":3,"guides-for-vercel-labs-agent-browser":409,"similar-k171hvn0e2t3jzczxm3vzspj3186mgj7-en":410},{"_creationTime":4,"_id":5,"children":6,"community":47,"display":48,"evaluation":52,"identity":249,"isFallback":245,"parentExtension":251,"providers":252,"relations":258,"repo":259,"tags":406,"workflow":407},1778698388090.9072,"k171hvn0e2t3jzczxm3vzspj3186mgj7",[7],{"_creationTime":8,"_id":9,"community":10,"display":12,"identity":17,"providers":22,"relations":40,"tags":42,"workflow":43},1778698388090.9075,"k175gma4kma3radvem5pa46wgn86nd1b",{"reviewCount":11},0,{"description":13,"installMethods":14,"name":15,"sourceUrl":16},"Automates browser interactions for web testing, form filling, screenshots, and data extraction",{"claudeCode":15},"agent-browser","https://github.com/vercel-labs/agent-browser",{"basePath":18,"githubOwner":19,"githubRepo":15,"locale":20,"slug":15,"type":21},"","vercel-labs","en","plugin",{"evaluate":23,"extract":35},{"promptVersionExtension":24,"promptVersionScoring":25,"score":26,"tags":27,"targetMarket":33,"tier":34},"3.0.0","4.4.0",98,[28,29,30,31,32],"browser-automation","web-testing","cli","automation","web-scraping","global","verified",{"commitSha":36,"plugin":37},"HEAD",{"mcpCount":11,"provider":38,"skillCount":39},"classify",1,{"parentExtensionId":5,"repoId":41},"kd74a6sfc26xa0ec2tpntj8r2986mn0a",[31,28,30,32,29],{"evaluatedAt":44,"extractAt":45,"updatedAt":46},1778698422911,1778698388091,1778698675646,{"reviewCount":11},{"description":49,"installMethods":50,"name":15,"sourceUrl":16},"Browser automation for AI agents",{"claudeCode":51},"vercel-labs/agent-browser",{"_creationTime":53,"_id":54,"extensionId":5,"locale":20,"result":55,"trustSignals":228,"workflow":247},1778698406247.6392,"kn79n25xvz76wfpbsh5zxbz00n86mmef",{"checks":56,"evaluatedAt":202,"extensionSummary":203,"features":204,"nonGoals":210,"promptVersionExtension":214,"promptVersionScoring":25,"purpose":215,"rationale":216,"score":217,"summary":218,"tags":219,"targetMarket":33,"tier":34,"useCases":223},[57,62,65,68,72,75,79,83,86,89,93,97,100,104,107,110,113,116,119,122,126,130,134,138,142,145,148,151,155,158,161,164,167,170,173,177,181,185,189,193,196,199],{"category":58,"check":59,"severity":60,"summary":61},"Practical Utility","Problem relevance","pass","The description 'Browser automation for AI agents' clearly states the user problem and the extension's purpose.",{"category":58,"check":63,"severity":60,"summary":64},"Unique selling proposition","The extension provides a specialized CLI for AI agents to automate browser interactions, which is a significant value beyond basic prompting.",{"category":58,"check":66,"severity":60,"summary":67},"Production readiness","The CLI offers a comprehensive set of commands for browser automation, including setup, core actions, advanced features like state management and debugging, indicating it's ready for real-world workflows.",{"category":69,"check":70,"severity":60,"summary":71},"Scope","Single responsibility principle","The extension focuses exclusively on browser automation and related tasks, without extending into unrelated domains.",{"category":69,"check":73,"severity":60,"summary":74},"Description quality","The displayed description 'Browser automation for AI agents' is concise, accurate, and reflects the extension's core functionality.",{"category":76,"check":77,"severity":60,"summary":78},"Invocation","Scoped tools","The extension exposes numerous specific tools like `click`, `fill`, `snapshot`, `screenshot`, which are narrow verb-noun specialists.",{"category":80,"check":81,"severity":60,"summary":82},"Documentation","Configuration & parameter reference","The README provides extensive documentation for all commands, options, and parameters, including usage examples and environment variable configurations.",{"category":69,"check":84,"severity":60,"summary":85},"Tool naming","Tool names are descriptive, verb-noun pairs within the browser automation domain (e.g., `click`, `fill`, `snapshot`).",{"category":69,"check":87,"severity":60,"summary":88},"Minimal I/O surface","Tools generally accept specific parameters (selectors, text, files) and return focused output, minimizing unnecessary data.",{"category":90,"check":91,"severity":60,"summary":92},"License","License usability","The extension is licensed under Apache-2.0, a permissive open-source license, and this is declared via a LICENSE file and a manifest.",{"category":94,"check":95,"severity":60,"summary":96},"Maintenance","Commit recency","The last commit was on May 7, 2026, which is recent, indicating active maintenance.",{"category":94,"check":98,"severity":60,"summary":99},"Dependency Management","The project uses `pnpm` for package management and has a lockfile (`pnpm-lock.yaml`), indicating good dependency management practices.",{"category":101,"check":102,"severity":60,"summary":103},"Security","Secret Management","The extension's documentation and features do not indicate it handles or echoes secrets; security features like an auth vault are opt-in and documented.",{"category":101,"check":105,"severity":60,"summary":106},"Injection","The documentation emphasizes secure practices, such as content boundary markers and domain allowlisting, and does not suggest executing untrusted code.",{"category":101,"check":108,"severity":60,"summary":109},"Transitive Supply-Chain Grenades","The extension bundles its dependencies and does not appear to fetch or execute remote code at runtime, mitigating supply chain risks.",{"category":101,"check":111,"severity":60,"summary":112},"Sandbox Isolation","The extension focuses on browser automation within its own controlled environment and does not show signs of attempting to modify files outside its scope.",{"category":101,"check":114,"severity":60,"summary":115},"Sandbox escape primitives","No detached process spawns or deny-retry loops were observed in the provided documentation or source code context.",{"category":101,"check":117,"severity":60,"summary":118},"Data Exfiltration","The extension's documentation highlights security features to prevent data exfiltration, and there are no indications of undocumented outbound calls.",{"category":101,"check":120,"severity":60,"summary":121},"Hidden Text Tricks","The README and documentation appear to be free of hidden text tricks or invisible characters used for steering models.",{"category":123,"check":124,"severity":60,"summary":125},"Hooks","Opaque code execution","The bundled Rust binary and provided documentation do not suggest obfuscated code or runtime script fetching.",{"category":127,"check":128,"severity":60,"summary":129},"Portability","Structural Assumption","The extension provides installation methods for various platforms (macOS, Linux, Windows) and does not appear to make assumptions about specific project structures outside its own bundle.",{"category":131,"check":132,"severity":60,"summary":133},"Trust","Issues Attention","With 142 issues opened and 155 closed in the last 90 days, the closure rate is approximately 52%, indicating active maintainer engagement.",{"category":135,"check":136,"severity":60,"summary":137},"Versioning","Release Management","The project has a meaningful semver version declared in the LICENSE file's appendix and a CHANGELOG.md, indicating clear release management.",{"category":139,"check":140,"severity":60,"summary":141},"Code Execution","Validation","While not explicitly detailed with a schema library, the extensive command-line interface with typed parameters and the focus on browser automation suggest a robust approach to input validation.",{"category":101,"check":143,"severity":60,"summary":144},"Unguarded Destructive Operations","Destructive operations are not a primary feature, and any potential for them would likely be guarded by explicit user commands or confirmations, aligning with safe practices.",{"category":139,"check":146,"severity":60,"summary":147},"Error Handling","The extensive CLI options and agent integration suggest well-defined error handling, with CLI commands expected to provide meaningful output and structured JSON for agents.",{"category":139,"check":149,"severity":60,"summary":150},"Logging","The extension provides debugging and observability features like `--debug`, console logging, and tracing, which facilitate auditing actions.",{"category":152,"check":153,"severity":60,"summary":154},"Compliance","GDPR","The extension automates browser interactions and does not inherently process personal data beyond what is present on web pages, with security features to manage data submission.",{"category":152,"check":156,"severity":60,"summary":157},"Target market","The extension provides tools for general web automation and does not contain any regional or jurisdictional logic, thus `targetMarket` is 'global'.",{"category":127,"check":159,"severity":60,"summary":160},"Runtime stability","The extension is a native Rust CLI with builds for multiple platforms and supports custom browser executables and cloud providers, ensuring broad runtime stability.",{"category":80,"check":162,"severity":60,"summary":163},"README","The README is comprehensive, well-structured, and clearly states the extension's purpose and capabilities.",{"category":69,"check":165,"severity":60,"summary":166},"Tool surface size","The extension exposes a rich set of tools (over 50 commands listed) but they are well-categorized and focused on browser automation, avoiding unnecessary bloat.",{"category":76,"check":168,"severity":60,"summary":169},"Overlapping near-synonym tools","While there are many commands, they represent distinct actions and are not near-synonyms that would cause ambiguity for an AI agent.",{"category":80,"check":171,"severity":60,"summary":172},"Phantom features","All advertised features, including advanced integrations and commands, appear to have corresponding implementations as detailed in the README and command list.",{"category":174,"check":175,"severity":60,"summary":176},"Install","Installation instruction","Installation instructions are provided for multiple methods (npm, Homebrew, Cargo, source) with clear copy-paste examples.",{"category":178,"check":179,"severity":60,"summary":180},"Errors","Actionable error messages","The CLI's design, including JSON output and detailed help, suggests that errors are actionable and informative for both users and agents.",{"category":182,"check":183,"severity":60,"summary":184},"Execution","Pinned dependencies","As a native Rust binary with npm packaging, dependencies are managed and pinned, ensuring reproducible builds.",{"category":69,"check":186,"severity":187,"summary":188},"Dry-run preview","not_applicable","The extension is primarily for interacting with a browser, and explicit destructive operations requiring a dry-run mode are not a core feature.",{"category":190,"check":191,"severity":60,"summary":192},"Protocol","Idempotent retry & timeouts","The CLI provides a default timeout and documentation on custom timeouts, supporting robust interaction patterns, and the underlying browser automation protocols likely handle state appropriately.",{"category":152,"check":194,"severity":60,"summary":195},"Telemetry opt-in","The extension's documentation emphasizes opt-in security features and does not mention any default telemetry collection.",{"category":69,"check":197,"severity":60,"summary":198},"Theme declaration","The marketplace description clearly states a theme: 'Browser automation for AI agents'.",{"category":135,"check":200,"severity":60,"summary":201},"Per-entry version metadata","The `marketplace.json` references the local source directory without a specific version, but the README's clear versioning and changelog provide reliable metadata.",1778698405672,"This is a command-line interface tool written in Rust that allows AI agents to automate browser interactions. It supports a wide range of actions like navigation, clicking, filling forms, taking screenshots, managing cookies, and advanced features like CDP connections and state persistence.",[205,206,207,208,209],"Automate browser actions (click, fill, type, navigate)","Support for CSS, text, and semantic locators","Generate snapshots and screenshots with optional annotations","Manage cookies, local storage, and network requests","Integrate with AI agents via JSON output and commands",[211,212,213],"Replacing web browsers for human use","Automating desktop applications outside the browser","Providing a graphical user interface for automation control","3.1.0","To provide AI agents with a powerful and reliable tool for interacting with web browsers and automating web-based tasks.","High quality implementation with extensive documentation and security features. Minor point on per-entry version metadata in marketplace.json.",99,"A robust and feature-rich browser automation CLI designed for AI agents.",[28,30,220,221,222],"webdriver","testing","ai-agent",[224,225,226,227],"Automating web testing and QA workflows","Performing data extraction from websites","Interacting with web applications for AI-driven tasks","Scraping website content and structure",{"codeQuality":229,"collectedAt":231,"documentation":232,"maintenance":235,"popularity":241,"security":243,"testCoverage":246},{"hasLockfile":230},true,1778698390136,{"descriptionLength":233,"readmeSize":234},32,61552,{"closedIssues90d":236,"forks":237,"hasChangelog":230,"openIssues90d":238,"pushedAt":239,"stars":240},155,2035,142,1778168700000,32896,{"npmDownloads":242},2851726,{"hasNpmPackage":230,"license":244,"smitheryVerified":245},"Apache-2.0",false,{"hasCi":230,"hasTests":230},{"updatedAt":248},1778698406247,{"basePath":18,"githubOwner":19,"githubRepo":15,"locale":20,"slug":15,"type":250},"marketplace",null,{"evaluate":253,"extract":255},{"promptVersionExtension":214,"promptVersionScoring":25,"score":217,"tags":254,"targetMarket":33,"tier":34},[28,30,220,221,222],{"commitSha":36,"marketplace":256,"plugin":257},{"name":15,"pluginCount":39},{"mcpCount":11,"provider":38,"skillCount":11},{"repoId":41},{"_creationTime":260,"_id":41,"identity":261,"providers":262,"workflow":401},1778698380039.9202,{"githubOwner":19,"githubRepo":15,"sourceUrl":16},{"classify":263,"discover":393,"extract":396,"github":397,"npm":400},{"commitSha":36,"extensions":264},[265,278,287,297,328,340,348,360,368,375,385],{"basePath":18,"description":49,"displayName":15,"installMethods":266,"rationale":267,"selectedPaths":268,"source":277,"sourceLanguage":20,"type":250},{"claudeCode":51},"marketplace.json at .claude-plugin/marketplace.json",[269,272,274],{"path":270,"priority":271},".claude-plugin/marketplace.json","mandatory",{"path":273,"priority":271},"README.md",{"path":275,"priority":276},"LICENSE","high","rule",{"basePath":18,"description":13,"displayName":15,"installMethods":279,"rationale":280,"selectedPaths":281,"source":277,"sourceLanguage":20,"type":21},{"claudeCode":15},"inline plugin source from marketplace.json at /",[282,283,284],{"path":273,"priority":271},{"path":275,"priority":276},{"path":285,"priority":286},"bin/agent-browser.js","low",{"basePath":288,"description":289,"displayName":290,"installMethods":291,"rationale":292,"selectedPaths":293,"source":277,"sourceLanguage":20,"type":296},"skill-data/agentcore","Run agent-browser on AWS Bedrock AgentCore cloud browsers. Use when the user wants to use AgentCore, run browser automation on AWS, use a cloud browser with AWS credentials, or needs a managed browser session backed by AWS infrastructure. Triggers include \"use agentcore\", \"run on AWS\", \"cloud browser with AWS\", \"bedrock browser\", \"agentcore session\", or any task requiring AWS-hosted browser automation.","agentcore",{"claudeCode":51},"SKILL.md frontmatter at skill-data/agentcore/SKILL.md",[294],{"path":295,"priority":271},"SKILL.md","skill",{"basePath":298,"description":299,"displayName":300,"installMethods":301,"rationale":302,"selectedPaths":303,"source":277,"sourceLanguage":20,"type":296},"skill-data/core","Core agent-browser usage guide. Read this before running any agent-browser commands. Covers the snapshot-and-ref workflow, navigating pages, interacting with elements (click, fill, type, select), extracting text and data, taking screenshots, managing tabs, handling forms and auth, waiting for content, running multiple browser sessions in parallel, and troubleshooting common failures. Use when the user asks to interact with a website, fill a form, click something, extract data, take a screenshot, log into a site, test a web app, or automate any browser task.","core",{"claudeCode":51},"SKILL.md frontmatter at skill-data/core/SKILL.md",[304,305,308,310,312,314,316,318,320,322,324,326],{"path":295,"priority":271},{"path":306,"priority":307},"references/authentication.md","medium",{"path":309,"priority":307},"references/commands.md",{"path":311,"priority":307},"references/profiling.md",{"path":313,"priority":307},"references/proxy-support.md",{"path":315,"priority":307},"references/session-management.md",{"path":317,"priority":307},"references/snapshot-refs.md",{"path":319,"priority":307},"references/trust-boundaries.md",{"path":321,"priority":307},"references/video-recording.md",{"path":323,"priority":286},"templates/authenticated-session.sh",{"path":325,"priority":286},"templates/capture-workflow.sh",{"path":327,"priority":286},"templates/form-automation.sh",{"basePath":329,"description":330,"displayName":331,"installMethods":332,"rationale":333,"selectedPaths":334,"source":277,"sourceLanguage":20,"type":296},"skill-data/dogfood","Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to \"dogfood\", \"QA\", \"exploratory test\", \"find issues\", \"bug hunt\", \"test this app/site/platform\", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.","dogfood",{"claudeCode":51},"SKILL.md frontmatter at skill-data/dogfood/SKILL.md",[335,336,338],{"path":295,"priority":271},{"path":337,"priority":307},"references/issue-taxonomy.md",{"path":339,"priority":286},"templates/dogfood-report-template.md",{"basePath":341,"description":342,"displayName":343,"installMethods":344,"rationale":345,"selectedPaths":346,"source":277,"sourceLanguage":20,"type":296},"skill-data/electron","Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include \"automate Slack app\", \"control VS Code\", \"interact with Discord app\", \"test this Electron app\", \"connect to desktop app\", or any task requiring automation of a native Electron application.","electron",{"claudeCode":51},"SKILL.md frontmatter at skill-data/electron/SKILL.md",[347],{"path":295,"priority":271},{"basePath":349,"description":350,"displayName":351,"installMethods":352,"rationale":353,"selectedPaths":354,"source":277,"sourceLanguage":20,"type":296},"skill-data/slack","Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include \"check my Slack\", \"what channels have unreads\", \"send a message to\", \"search Slack for\", \"extract from Slack\", \"find who said\", or any task requiring programmatic Slack interaction.","slack",{"claudeCode":51},"SKILL.md frontmatter at skill-data/slack/SKILL.md",[355,356,358],{"path":295,"priority":271},{"path":357,"priority":307},"references/slack-tasks.md",{"path":359,"priority":286},"templates/slack-report-template.md",{"basePath":361,"description":362,"displayName":363,"installMethods":364,"rationale":365,"selectedPaths":366,"source":277,"sourceLanguage":20,"type":296},"skill-data/vercel-sandbox","Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include \"Vercel Sandbox browser\", \"microVM Chrome\", \"agent-browser in sandbox\", \"browser automation on Vercel\", or any task requiring Chrome in a Vercel Sandbox.","vercel-sandbox",{"claudeCode":51},"SKILL.md frontmatter at skill-data/vercel-sandbox/SKILL.md",[367],{"path":295,"priority":271},{"basePath":369,"description":370,"displayName":15,"installMethods":371,"rationale":372,"selectedPaths":373,"source":277,"sourceLanguage":20,"type":296},"skills/agent-browser","Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to \"open a website\", \"fill out a form\", \"click a button\", \"take a screenshot\", \"scrape data from a page\", \"test this web app\", \"login to a site\", \"automate browser actions\", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.",{"claudeCode":51},"SKILL.md frontmatter at skills/agent-browser/SKILL.md",[374],{"path":295,"priority":271},{"basePath":18,"description":376,"displayName":15,"installMethods":377,"license":244,"rationale":378,"selectedPaths":379,"source":277,"sourceLanguage":20,"type":30},"Browser automation CLI for AI agents",{"npm":15},"cli ecosystem detected at /",[380,382,383,384],{"path":381,"priority":271},"package.json",{"path":273,"priority":271},{"path":275,"priority":276},{"path":285,"priority":307},{"basePath":30,"installMethods":386,"rationale":387,"selectedPaths":388,"source":277,"sourceLanguage":20,"type":30},{"cargo":15},"cli ecosystem detected at cli",[389,391],{"path":390,"priority":271},"Cargo.toml",{"path":392,"priority":307},"src/main.rs",{"sources":394},[395],"manual",{"npmPackage":15},{"closedIssues90d":236,"description":376,"forks":237,"homepage":398,"license":244,"openIssues90d":238,"pushedAt":239,"readmeSize":234,"stars":240,"topics":399},"https://agent-browser.dev",[],{"downloads":242},{"classifiedAt":402,"discoverAt":403,"extractAt":404,"githubAt":404,"npmAt":405,"updatedAt":402},1778698387373,1778698380040,1778698383100,1778698385198,[222,28,30,221,220],{"evaluatedAt":248,"extractAt":45,"updatedAt":408},1778698675865,[],[]]