reverse-engineer

A method for discovering and reproducing a website’s hidden data interfaces, including APIs, live WebSocket connections, and encrypted responses. It moves through increasingly advanced investigation steps as needed.

In plain words
What is it for?
Use it when ordinary page scraping cannot find the required data. It covers network capture, request replay, token discovery, JavaScript analysis, decryption, WebSockets, GraphQL, and binary formats.
Why use it?
Important data may not appear in the page HTML or may be protected behind undocumented requests and encoded responses. The method helps identify how the browser obtains and decodes that data.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agentcomputerai/torch/reverse-engineer
Any agent
npx skills add AgentComputerAI/torch --skill reverse-engineer
Clone the repo
git clone --depth 1 https://github.com/AgentComputerAI/torch

Made for: Claude Code, Codex.

Per session 96 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,361 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00096 $0.07361
Opus 5 $0.00048 $0.03681
Sonnet 5 $0.00019 $0.01472
Haiku 4.5 $0.00010 $0.00736

Measured yesterday against content hash dcfaa1daf05e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

reverse-engineer scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

**When**: Phase 0 curl returns HTML with no target data. The data loads via XHR/fetch after JS runs.
skills/reverse-engineer/SKILL.md · 675 lines

How it starts

The opening of the file, as written. The whole thing — 675 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Reverse Engineering

When the data isn't in the HTML, it's in an API. When the API isn't public, it's hidden in the JavaScript. When the JavaScript is obfuscated, the keys are still in the page source — because the browser needs them to decrypt, which means you can too.

This skill is an escalation ladder. Start at Level 1 and stop as soon as you have what you need.

Escalation ladder

  Level 1   📡  Network capture          watch DevTools, find the API call
  Level 2   🔬  Response classification  Shannon entropy — plaintext vs compressed vs encrypted
  Level 3   🔁  API replay               copy as cURL, replay with fetch, 403 bypass tricks
  Level 4   🔑  Token extraction         find auth tokens, CSRF, API keys in page source
  Level 5   📦  JS deobfuscation         unpack webpack, deobfuscate, read the source
  Level 6   🔐  Payload decryption       NaCl / AES-GCM / CryptoJS — extract keys, decrypt
  Level 7   🔌  WebSocket interception    establish WS, subscribe to channels, decode frames
  Level 8   🔮  GraphQL reconstruction    extract queries from JS, force PersistedQueryNotFound
  Level 9   🧬  Protobuf decoding        reverse-engineer .proto schema from binary blobs

Level 1 — Network capture

When: Phase 0 curl returns HTML with no target data. The data loads via XHR/fetch after JS runs.

Open the page in the real Chrome debug port, capture every API call during page load:

const page = await browser.newPage();

const apiCalls = [];
page.on("response", async (res) => {
  const url = res.url();
  const ct = res.headers()["content-type"] || "";
  if (
    ct.includes("json") || ct.includes("grpc") || ct.includes("protobuf") ||
    url.includes("/api/") || url.includes("/graphql") || url.includes("/v1/") || url.includes("/v2/")
  ) {
    try {
      const body = await res.text();
      apiCalls.push({ url, status: res.status(), contentType: ct, size: body.length, sample: body.slice(0, 500) });
    } catch {}
  }
});

await page.goto(url, { waitUntil: "networkidle2" });

for (const call of apiCalls) {
  console.log(`${call.status} ${call.url} (${call.contentType}, ${call.size}B)`);
  console.log(`  ${call.sample.slice(0, 200)}`);
}

Read the full file on GitHub · 675 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 675 lines · 96 tokens per session scan A dcfaa1daf05e

Subscribe to this mod's changes

reverse-engineer is a skill published in the GitHub repository AgentComputerAI/torch (5 stars, last pushed 4mo ago), licensed MIT. It adds 96 tokens to every session and 7,361 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

tiny-web-crawler

Crawl from one or more starting web pages, fetch readable content, search within pages, follow relevant links, and stop when the requested information is found or a bounded limit is reached.

leon-ai/leon · 42 tokens

playwright-cli

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

VoltAgent/voltagent · 52 tokens

browserstack

../../../engineering-team/playwright-pro/skills/browserstack/SKILL.md.

alirezarezvani/claude-skills · 0 tokens

analyze-performance-traces

Analyze Chrome, Chromium, Electron, React DevTools, or Perfetto-compatible JSON traces and audit user-reported profiling findings without loading large artifacts into context; prove trigger-to-render/layout chains, separate measured facts from source inference, find exact code choke points, classify forced layout and…

tutti-os/tutti · 127 tokens

feature-demo-recording

Record a demo video of a web feature from a real browser. Two modes -- a NARRATED film where measured voiceover drives the timeline (designed slides, subtitles, punch-in camera, rendered from an HTML timeline), and a SILENT evidence clip for a PR or a QA pass. Use when the user asks to record a video, demo, or screen…

kirodotdev/KiroCrew · 90 tokens

browser-recording

Record a browser flow as a video/GIF for evidence — animations, transitions, and multi-step interactions that a still screenshot cannot prove. Drives the project's own Playwright through a bundled runner, then converts to mp4 + GIF via ffmpeg. Use when the user asks to record a demo, capture a GIF or video of the UI…

kirodotdev/KiroCrew · 85 tokens