More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    JS Reverse: Rebuild Browser JavaScript from Evidence

    1

    Reverse-engineer browser JavaScript the evidence-first way: watch, sample the smallest reproduction, rebuild locally, and converge.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    JS Reverse: Rebuild Browser JavaScript from Evidence

    JS Reverse: Rebuild Browser JavaScript from Evidence

    Example session with this skill installed

    A login page sends sign=9f2ac1... with each request. Reproduce the signing routine locally so I can call the API from my own integration.

    • Read your context and instructions
    • Compiled the js reverse

    Watch: the network log showed POST /api/login with sign in the body; the initiator chain pointed to app.3f2a.js, function h() called from login(). Evidence note written: URL, initiator chain, script URL, candidate function. Sample: an XHR breakpoint fired on the login request; the captured body showed username, password, timestamp, and sign. Evaluating h.toString() showed it calls md5(username + password + t + salt). The salt lookup read window.__SALT — captured exactly, never invented. Rebuild: copied h and the md5 implementation into a local Node script, stubbed window.__SALT with the captured string, and nothing else — no full DOM, no polyfills, no guessed environment. Converge: fed the captured username, password, and timestamp; the local sign matched byte for byte. Re-ran with a fresh timestamp against the real endpoint and the request was accepted. Deep-dive skipped deliberately: this is a one-off signature for the user's own integration, so convergence was the finish line, not deobfuscation. Divergence log recorded every patch and the observed behavior that justified it, so the reversal can be re-run and reviewed by anyone who was not in the room. No deobfuscation was needed because convergence held on fresh samples.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Frontend reversal fails when people guess. This skill keeps every step anchored to something observed in a real browser session: watch the network traffic, capture the smallest runtime sample that reproduces a signing or encryption routine, rebuild it locally with a minimal environment, then drive it to convergence. It covers the full loop — watch, sample, rebuild, converge, and deep-dive only when the algorithm will be reused — plus capture discipline, minimal-environment rules, and deobfuscation triage for string arrays, control-flow flattening, anti-debugger checks, and VM-based obfuscation. The guardrail is explicit: this is for authorized analysis only — your own integrations, security research you have permission for, or compatibility work.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.

    Authorized analysis only: your own integrations, security research you have permission for, or compatibility work. It is not a tool for bypassing access controls on services you are not entitled to use.
    VM-based obfuscation is treated as its own project — the skill does not pretend hand-reading it is a realistic step.
    Convergence on captured samples is not proof against future server-side changes; re-verify with fresh samples when the target changes.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Watching. Open the instrumented browser, find the request you care about, and trace it back to the function. Reading obfuscated code cold is the slowest possible start.

    As little as the evidence demands. Add a stub only after an observed access or a runtime error requires it, and stub the narrowest behavior — the exact captured value, not a general polyfill.

    When the local copy matches the page on fresh samples, not just the originals. If the algorithm will be reused long-term, that is when you do the real deobfuscation work — not before.

    New in Security