AI Browser Agent buildup

We're building a custom browser-based AI automation tool powered by Anthropic's Claude API. This is a fully custom build using Claude's Agent SDK and computer-use/browser-use tools — not a rebrand or repackaging of any existing browser extension. We're looking for a developer to design, build, and ship this from the ground up. What You'll Do ● Architect and build a browser automation agent using the Claude API's computer use and/or browser use tools ● Implement the agent loop (tool call → execute action → return result → repeat) that drives Claude's interaction with a live browser session ● Stand up a sandboxed execution environment (containerized browser instance/VM) for the agent to operate in safely ● Build the front-end/UI layer for triggering, monitoring, and reviewing agent runs ● Implement guardrails: domain allowlisting, human-confirmation checkpoints for consequential actions, and prompt-injection safeguards ● Handle authentication, session management, and deployment (hosting, CI/CD) ● Write documentation for setup, configuration, and internal usage Requirements ● Strong experience with Python or TypeScript/Node.js ● Experience integrating LLM APIs (Anthropic, OpenAI, or similar) into production applications ● Familiarity with browser automation tooling (Playwright, Puppeteer, or similar) ● Comfortable working with containerized environments (Docker) for isolated execution ● Understanding of API security best practices (credential handling, sandboxing, least-privilege access) Nice to Have ● Prior experience with Anthropic's Claude API, Claude Agent SDK, or the computer-use/browser-use tool family ● Experience building agentic/autonomous AI workflows ● Front-end experience (React or similar) for building a monitoring/control dashboard