Anthropic has a feature in Claude Code literally called "undercover mode." No off switch. It quietl

2026-05-04 · Source: What's AI by Louis-François Bouchard · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Novice, quick

Summary

Anthropic's Claude AI features an "Undercover mode," implemented in 90 lines of `undercover.ts` code, which prevents the AI from disclosing its nature. This mode also removes "co-authored by" attributions from GitHub commits and automatically activates for Anthropic employees contributing to external repositories, without an off switch. The internal code explicitly states, "You are operating undercover. Do not blow your cover." This feature has raised concerns given Anthropic's public brand emphasis on transparency and AI safety, despite similar functionalities likely existing within other companies.

Key takeaway

For executives overseeing AI product development and brand messaging, you should audit internal tools for features that might contradict public transparency commitments. Ensure that your company's operational practices, such as AI identity concealment or attribution policies, align with your stated values to maintain trust and avoid reputational damage.

Key insights

Anthropic's "Undercover mode" conceals Claude's AI identity and removes attribution for internal contributors.

Principles

• AI identity concealment is a design choice.
• Transparency claims can conflict with internal features.

In practice

• Review code for hidden AI identity features.
• Assess brand messaging against internal tooling.

Topics

• Undercover Mode
• Claude Code
• AI Identity Concealment
• GitHub Attribution
• Anthropic Transparency

Best for: CTO, VP of Engineering/Data, Executive, AI Ethicist, Tech Journalist, Director of AI/ML

Related on AIssential

• Here's what that Claude Code source leak reveals about Anthropic's plans — Anthropic's leaked Claude Code reveals future features like persistent agents, memory consolidation, and anonymous open-source contributions.
• Claude’s code: Anthropic leaks source code for AI software engineering tool — Human error led to a significant source code leak for Anthropic's Claude Code, raising security and competitive concerns.
• 🗞️ Anthropic launched Claude Code Security to scan your code repositories for bugs and suggest security patches. — AI is advancing in code security, robotics platforms, and developer tools, yet faces challenges in user adoption and intellectual property protection.
• Claude Code source code LEAKED — An accidental source code leak revealed Anthropic's Claude development roadmap and future AI capabilities.
• the claude leak made one thing harder to ignore — Agent products are sophisticated control systems, with orchestration and explicit control layers being critical.
• Post-Mortem of Anthropic's Claude Code Leak — AI system IP resides in the agent harness, not just model weights, driving agentic development innovation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by What's AI by Louis-François Bouchard.