MMCPMarkets
XW

xscanzm/wechat-4x-decrypt

Developer tools

This page was translated automatically. Technical details may differ from the original. View the original

24 stars Quality 57 Trend 19MIT
Repository Install

WeChat 4.x Local Chat Record Decryption and Retrieval Skill (Windows / Zero Third-Party Dependencies). Produced by: Skill · Monday · Pipi Future

Overview

A skill for AI Agents such as WorkBuddy / Claude Code / Codex is used to decrypt and retrieve (Weixin.exe / xwechat_files / db_storage) local chat databases on **, thereby exporting, searching, and summarizing your own chat records. WeChat 4.x stores chat records in the locally encrypted SQLite database (WCDB / SQLCipher variant). This skill provides a pure scripting workflow (using only Windows' built-in bcrypt.dll + Python standard library), enabling AI Agents to: 1. Automatically detects the currently running WeChat version, process, data directory, and active accounts; 2. Extracts decryption keys from **; 3. Batch decrypts encryption databases into plaintext SQLite; 4. Searches, exports, and summarizes chat content by keyword + timeframe. ⚠️ 。 For WeChat 3.x (WeChat.exe), please switch to pywxdump. ⚠️ 。 Key extraction depends on Windows CNG (bcrypt.dll). : Only operate on the user's local machine, the user's own account, and when the user specifically requests it; Do not read conversations unrelated to tasks, do not send chat content to third-party services, and do not retain unrelated personal privacy information in output. - WeChat 4.1+ process memory**, only one WCDB passphrase object; - Key tool locates com. Tencent.WCDB.Config.Cipher string object → Retrieve blob → XOR along the object chain to decode (27-byte mask) → obtain x'' → HMAC-SHA512 validation; - Decryption algorithm is SQLCipher 4 variant (AES-256-CBC + HMAC-SHA512, page_size=4096, reserve=80); - After decryption, the acontent column in the message_fts_v4_* table of message_fts.db is already the decoded text; using the name2id table can restore the speaker and session name. Therefore, any old scheme that relies on "searching for x'...' in memory or 'taking salt adjacent 32 bytes'" (pywxdump / NTLx-wechat-decrypt / ZedeX scripts) will inevitably hit zero on version 4.1+—this skill is not included. Copy the entire directory of this repository to the skills directory of the agent (directory name is the skill name):

Overview (machine translated)

Produced by skill · Monday · Pipi Future

A skill for AI Agents such as WorkBuddy, Claude Code, Codex, etc., used to decrypt and retrieve the local chat database of WeChat 4.x (Weixin.exe / xwechat_files / db_storage) on Windows native, thereby exporting, searching, and summarizing your own chat records.

1. What is this?

WeChat 4.x stores chat records in the locally encrypted SQLite database (WCDB / SQLCipher variants). This skill provides a pure scripting workflow with zero third-party dependencies (using only Windows’ built-in bcrypt.dll + Python standard library), enabling AI Agents to:

  1. Automatically detects the currently running WeChat version, processes, data directory, and active accounts;
  2. Extract the decryption key from the memory of the running WeChat process;
  3. Batch decrypt the encryption library into plaintext SQLite;
  4. Search, export, and summarize chat content by keyword + time window.

⚠️ > Only supports WeChat 4.x (Weixin.exe). For WeChat 3.x (WeChat.exe), please switch to pywxdump. ⚠️ > Supports Windows only. Key extraction depends on Windows CNG(bcrypt.dll).

2. Applicable / Not Applicable

Applicable scenarios Not applicable
Export your WeChat chat history for backup WeChat 3.x (using pywxdump)
Locally retrieve a specific conversation or group Non-Windows system
Let AI help you summarize/summarize what you talked about during a certain period Read other people’s accounts and your logged-in status
Analyze recent discussion hotspots in a certain group Online/cloud WeChat data

Compliance: Only operate on the user’s local machine, the user’s own account, and when explicitly requested by the user; Do not read conversations unrelated to the task, do not transmit chat content to third-party services, and do not retain unrelated personal privacy information in output.

3. Working Principle (30-Second Version)

  • WeChat 4.1+ process memory does not have a plaintext key, only a WCDB passphrase object;
  • Key tool locates com.Tencent.WCDB.Config.Cipher string object → Retrieves blob → XOR along the object chain to decode (27-byte mask) → obtain x' ' → HMAC-SHA512 validation;
  • The decryption algorithm is a variant of SQLCipher 4 (AES-256-CBC + HMAC-SHA512, page_size=4096, reserve=80);
  • After decryption, the acontent column in the message_fts_v4_* table of message_fts.db is already the decoded text; using the name2id table can restore the speaker and session name.

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add xscanzm/wechat-4x-decrypt
New to MCP? Start with the mcpskill.org guide →