Show HN: I found API keys in my coding agent history so I built Agent Scrub (github.com)

A developer recently introduced Agent Scrub, a tool designed to enhance security by scanning and redacting API keys and other sensitive information inadvertently stored in the history files of various AI coding agents like Claude Code and GitHub Copilot. As these tools often save conversation histories that may include plaintext secrets, Agent Scrub operates locally to identify and redact these potential vulnerabilities without uploading any data. This ensures that developers can work confidently, knowing that their local history does not expose sensitive credentials. The significance of Agent Scrub lies in its ability to mitigate security risks associated with the widespread use of AI coding tools. It supports multiple platforms and can detect and manage secrets across various coding agents, ensuring that sensitive information remains protected. Agent Scrub allows users to reveal, redact, or automate the redaction of detected secrets, thereby streamlining the security process. Additionally, it handles permissions cautiously, only accessing necessary local files without disturbing active credentials, making it a valuable addition to the toolkit of developers concerned about data leakage in an increasingly AI-driven coding environment.