UnkL4b/GitMiner3 — reverse-engineered prompt

Reverse engineered prompt

Build me a Python tool that can search public GitHub repositories for possible exposed secrets using GitHub search queries, download the matching files, scan them with configurable regex patterns, and save the results in a way that is easy to review later.

I want to be able to give it either one search query or a text file with several queries. It should use a GitHub personal access token, support simple config files for paths and detection rules, keep a SQLite history of what it searched and what it found, and optionally export a CSV summary. Please also generate a readable Markdown report for each query with a short summary, severity levels, technical details, and links or references to the source files.

Keep it modular and clean so the scanning, config loading, GitHub API access, file handling, database work, and report writing are separated. Include sensible error handling and rate limit handling. Assume Python 3.11. If anything is unclear, check current GitHub API docs online and make it work end to end from the command line.

Want more depth? Deep Reverse