This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
AI-assisted skills, rules, and prompts for reverse engineering (Web/JS, Android/iOS App). This is a prompt/knowledge repository, not a runnable application — there are no build steps, tests, or dependencies to install.
skills/— Task-oriented Claude Code skills (SKILL.md files with frontmatter). Each skill is a self-contained workflow for a specific reverse engineering task.rules/— Background knowledge documents providing reference guides, tool usage patterns, and domain expertise for reverse engineering.prompts/— Common prompt templates (planned, not yet populated).
This toolkit is designed to work with the js-reverse MCP server for browser debugging. The MCP provides Chrome DevTools Protocol access: network inspection, breakpoints, script source retrieval, function hooking, and JS evaluation.
- Skills use
$ARGUMENTSas placeholder for user-provided parameters - Analysis progress is tracked in
PROGRESS.md(created per-project, not in this repo) - Deobfuscation scripts go in
scripts/deobfuscate_{target}.js - Deobfuscated output goes in
source/deobfuscated/{filename}_deobf.js - Python algorithm implementations go in
output/sign.py - Content is written in Chinese (Simplified)
- Use
uvto run Python and manage Python projects
Skills are defined as SKILL.md files with YAML frontmatter (name, description, argument-hint). Follow the pattern in skills/find-crypto-entry/SKILL.md: define clear step-by-step workflows with decision points and fallback paths between steps.
Rules are standalone markdown files in rules/. They should provide structured reference material (identification tables, tool command references, common patterns) rather than step-by-step workflows.