Skip to main content

jsrecon

Master JavaScript reconnaissance, de-minification, source map unpacking, deep secret mining, and endpoint discovery skill for bug bounty hunting based on TBHM and modern offensive JS pipelines.

معلومات المصدر

المستودع
pruthuraut/BB-skill
آخر نشاط في المصدر
٥ أكتوبر ٢٠٢٦ في ١٧:٣١
لغة SKILL.md المكتشفة
الإنجليزية
النجوم
٠
التفرعات
٠

خيارات التثبيت

يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.

مراجعة ملفات المصدر

اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.

مستكشف الملفات
11 ملفات

عرض SKILL.md

SKILL.md
تعليمات المصدر · معاينة للقراءة فقط
name
jsrecon
description
Master JavaScript reconnaissance, de-minification, source map unpacking, deep secret mining, and endpoint discovery skill for bug bounty hunting based on TBHM and modern offensive JS pipelines.
version
1.0.0
author
Bug Bounty Multi-Agent Skills Framework
compatibility
["Antigravity / Gemini CLI","Claude Code","Cursor IDE","OpenAI Codex / OpenCode","DeepSeek"]
# `jsrecon` — Master Bug Bounty JavaScript Reconnaissance Skill ## Overview `jsrecon` is an end-to-end, multi-agent offensive JavaScript reconnaissance, de-minification, and source map recovery skill. It operates in **Zero-Redundancy Mode** by automatically inheriting previously discovered assets from [`subdomainenum`](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/subdomainenum/SKILL.md) (`live_subdomains.txt`) and [`contentdiscovery`](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/contentdiscovery/SKILL.md) / [`linkparamdiscovery`](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/linkparamdiscovery/SKILL.md) (`all_discovered_urls.txt`, `js_files.txt`), only performing gap-filling crawling when running standalone. It formats dense code with `js-beautify`, unpacks original TypeScript/React source trees from `.js.map` files, runs a 60+ secret regex catalog alongside TruffleHog, and extracts all internal API routes, body keys, and query parameters. ``` [ PRIOR ARTIFACTS: live_subdomains.txt, all_discovered_urls.txt, js_files.txt ] │ (Zero-Redundancy Reuse) │ ▼ ┌───────────────────────────────────────────────┐ │ Subagent 01: Artifact Ingestion & Gap-Fill │ │ (Inherit prior crawl, live probe via HTTPx, │ │ parallel download into artifacts/js_raw/) │ └───────────────────────┬───────────────────────┘ │ ┌───────────────┴───────────────┐ ▼ ▼ [ live_js_urls.txt ] [ data_endpoints.txt ] │ (.json & .xml routes) ▼ ┌─────────────────────────────────────────────────┐ │ Subagent 02: Beautification & Source Map Unpack │ │ (js-beautify, .js.map probe, unminify TS source)│ └────────────────────────┬────────────────────────┘ │ ┌───────────────┴───────────────┐ ▼ ▼ ┌─────────────────────────────┐ ┌─────────────────────────────┐ │ Subagent 03: Secrets & TH │ │ Subagent 04: Deep Endpoints │ │ (60+ Regex Suite, TruffleHog│ │ (6-Way Endpoint Extraction, │ │ Verified Provider Scanning) │ │ Parameters, Router Tables) │ └──────────────┬──────────────┘ └──────────────┬──────────────┘ │ │ └───────────────┬───────────────┘ ▼ ┌─────────────────────────────────────────────────┐ │ Subagent 05: Runtime Hooking & Alerting │ │ (fetch/XHR console hooks, DOM sinks, Slack alert│ └────────────────────────┬────────────────────────┘ ▼ [ CONSOLIDATED ARTIFACTS ] • live_js_urls.txt • artifacts/js_recon/source_unpacked/ (Raw Source) • secrets_findings.txt & trufflehog_verified.txt • ALL_FINAL_ENDPOINTS.txt • parameters_discovered.json ``` --- ## The 5 Specialized Subagents 1. **[Subagent 01: Multi-Source JavaScript Harvesting, Extension Sorting & Parallel Downloads](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/subagents/subagent_01_harvest_download.md)** *Focus:* Wildcard subdomain resolution, active crawling (`katana`, `hakrawler`), historical scraping (`gau`, `waybackurls`), deterministic extension sorting (`.js`, `.json`, `.xml`, parameters), live verification via `httpx`, and content-hashed local caching. 2. **[Subagent 02: Automated De-Minification & Source Map (.map) Reconstruction](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/subagents/subagent_02_beautify_sourcemap.md)** *Focus:* Formatting with `js-beautify`, source map detection (`//# sourceMappingURL=`, `SourceMap:` headers, direct `.map` probing), unpacking complete developer source trees into `source_unpacked/`, and harvesting `TODO`/`FIXME` notes. 3. **[Subagent 03: High-Fidelity Secrets, Token Mining & TruffleHog Verification](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/subagents/subagent_03_secrets_trufflehog.md)** *Focus:* 60+ regex detection catalog (AWS, GCP, Firebase, Stripe, Slack, Discord, JWT, GitHub, Twilio, OpenAI, Anthropic, Private Keys) and live cryptographic secret verification using TruffleHog. 4. **[Subagent 04: Deep Endpoint Extraction, Parameter Mining & Architectural Analysis](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/subagents/subagent_04_endpoints_params.md)** *Focus:* The 6-Way endpoint extraction suite (Absolute, Keywords, Dynamic `${id}`, Raw URLs, React/Vue Router routes, and CDN cleaning), body parameter mining (`userId`, `role`, `isAdmin`), and query string harvesting. 5. **[Subagent 05: Dynamic Runtime Hooking, DOM Sinks & Notification Dispatch](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/subagents/subagent_05_runtime_dom_alerting.md)** *Focus:* In-browser console monkeypatching for `window.fetch()` and `XMLHttpRequest`, client-side DOM XSS sink auditing (`innerHTML`, `eval()`, `dangerouslySetInnerHTML`), postMessage listeners, and Slack/Discord alerting via `slackcat`. --- ## Standalone Pipeline Scripts - **[jsx_pipeline.sh](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/scripts/jsx_pipeline.sh)** — High-throughput parallel bash pipeline supporting multi-target harvesting, beautification, 60+ pattern scanning, and TruffleHog. - **[js_analyzer.py](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/scripts/js_analyzer.py)** — Cross-platform Python script for AST extraction, source map unpacking, and endpoint classification. --- ## Reference Guides - **[Master JavaScript Secret Regex Catalog](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/references/secret_regex_catalog.md)** - **[Endpoint Analysis & Triage Matrix](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/references/endpoint_analysis_matrix.md)** - **[Browser Console Snippets for Dynamic Recon](file:///c:/Users/rautp/Documents/BBskill/.agents/skills/jsrecon/references/browser_console_snippets.md)** --- ## Standard Output Artifacts Each execution writes to `artifacts/js_recon/<domain>/`: - `urls_all.txt` & `live_js_urls.txt` — Harvested and live-verified JavaScript URLs. - `data_endpoints_all.txt` — Isolated `.json` and `.xml` API endpoints. - `parameterized_urls_all.txt` — Endpoints containing query parameters. - `js_raw/` & `js_beautified/` — Locally cached and de-minified JavaScript files. - `source_unpacked/` — Reconstructed original developer source trees unpacked from `.js.map`. - `secrets_findings.txt` — Matched credentials and high-value tokens. - `tools/trufflehog_verified_findings.txt` — Cryptographically verified active credentials. - `ALL_FINAL_ENDPOINTS.txt` — Deduplicated, CDN-cleaned application endpoints. - `parameters_body.txt` & `parameters_query.txt` — Extracted request parameters.
عرض على GitHub