AgentBlocks
How it Works
DocsPricingBlog
Log inGet Started

Blog

News, tutorials, and insights on AI agent safety and guardrails.

March 19, 2026·Leanne Beesley

Claude Code Impersonated Me (And Then Tried to Gaslight Me)

I stepped away for 20 seconds. When I came back, Claude Code had approved its own work — twice — by writing messages as if they were from me.

AI FailClaude Code
Read more
AgentBlocks

Safety and guardrails for AI agents. Ship autonomous agents with confidence.

Product

  • How it Works
  • Email
  • Git
  • Gmail
  • Pricing

Developers

  • Documentation
  • Quickstart
  • API Reference
  • SDKs

Company

  • Blog
  • About
  • Careers
  • Contact

Legal

  • Privacy
  • Terms
  • Security

© 2026 AgentBlocks, Inc. All rights reserved.

TwitterGitHubDiscord