Skip to content
Assist Guild AG markAGAssistGuild
Model releases

Apiiro report links 1,122 exposed GitHub secrets to Claude Code’s Don’t Ask Again feature

An external analysis associates public credential leaks with a specific preference setting in Anthropic’s coding agent.

Key points

  • Apiiro reported that 1,122 public GitHub files contain secrets linked to the Don’t Ask Again feature.
  • Hacker News tracked the report with 162 points and 111 comments.
  • The source does not confirm whether Anthropic acknowledged the specific link or provided a patch.
Apiiro report links 1,122 exposed GitHub secrets to Claude Code’s Don’t Ask Again feature
Depth
6 outlets · newest source 2h ago

An analysis published on apiiro.com associates 1,122 publicly accessible GitHub files with credentials exposed through Claude Code’s Don’t Ask Again feature.

Evidence: confirmedClaude API Release Notes page updated: We’ve launched Claude Fable 5.2 (claude-fable-5-2), built for demanding reasoning and long

pagewatch-anthropic

The claim

The response

The source material provides no direct statement from Anthropic regarding this specific finding. Hacker News aggregated the report, generating 162 points and 111 comments, but the content of those discussions is not detailed in the provided metadata.

Evidence: reportedClaude Code's "Don't Ask Again" exposes passwords in 1,122 public GitHub files

aifleet-hn-front

The practical upshot

Product teams must verify if their agents persist sensitive data in ways that users might inadvertently commit.

Claims versus evidence

The evidence consists of the report’s identification of these files. The source does not explain the technical mechanism of how the files were committed, only that the feature is linked to the exposure. Uncertainty remains regarding the exact workflow that led to these commits.

Why it matters

  • The Don’t Ask Again feature is linked to secret exposure in public GitHub files according to the report.
  • Passwords were exposed in 1,122 public GitHub files due to this vulnerability.
  • Teams using Claude Code must audit their repository history for inadvertently committed secrets linked to agent state files.
  • The issue highlights risks associated with coding agents leaking secrets through public repositories.

Topics: Claude Code, Anthropic, GitHub, secrets management, AI security

Updated

Story timeline

  1. 07 Nov 2026 Claude Code's "Don't Ask Again" exposes passwords in 1,122 public GitHub files
  2. 09 Nov 2026 Claude Fable 5.2 and Mythos 5.2
  3. 09 Nov 2026 Claude API Release Notes page updated: We’ve launched Claude Fable 5.2 (claude-fable-5-2), built for demanding reasoning and long
  4. 09 Nov 2026 Anthropic News page updated: Introducing Claude Fable 5.2 and Claude Mythos 5.2
  5. 09 Nov 2026 Anthropic says during prerelease testing Mythos 5.2 and Fable 5.2 formalized the proof of Fermat's Last Theorem and solved two open black hole physics problems (Jay Peters/The Verge)
  6. 09 Nov 2026 v2.1.321
  7. 09 Nov 2026 Meta launches Muse Ember, a small AI model built for the repetitive work agents actually do
  8. 09 Nov 2026 Claude API Release Notes page updated: We’ve launched Claude Fable 5.2 (claude-fable-5-2), built for demanding reasoning and long
  9. 09 Nov 2026 Anthropic prices Fable 5.2 and Mythos 5.2 at $8/1M input and $40/1M output tokens, down 20% from 5.1; fast mode costs $16/1M input and $80/1M output tokens (Frederic Lardinois/The New Stack)

The strongest doubt

Counterpoint from the sources

The source does not specify whether the secrets were actively used or merely present in the files.

Ask the article

Local keyword search of this article; no generated answers.

Which of these is confirmed?

Check the answer

An analysis published on apiiro.com associates 1,122 publicly accessible GitHub files with credentials exposed through Claude Code’s Don’t Ask Again feature.

The matched primary source and announcement wording support the confirmed label.

Did this article give you the evidence you needed?

No vote saved on this device.
Your choice stays on this device. No public tally.

Terms in this story

Don’t Ask Again

A feature in Claude Code linked to exposed passwords in the report.

GitHub

A web-based platform for version control and collaboration using Git.

Questions readers ask

How many files are affected by this report?

The apiiro.com report identifies 1,122 public GitHub files linked to the exposure.

Does Anthropic confirm this issue?

The provided source material does not contain a confirmation or statement from Anthropic.

What feature is linked to the leaks?

The Don’t Ask Again feature in Claude Code is linked to the exposed secrets.

Connected stories

This storyClaude Code v2.1.314 restricts file writes and compacts idle long conversationsClaude Code v2.1.314 restricts fOpenAI and Anthropic seek smaller data centers as local bans mountOpenAI and Anthropic seek smalleBinance Heist Involved Malware, Fake Interviews and Internal AI ToolBinance Heist Involved Malware, OpenAI and Anthropic Pledge to Shield Midterm Elections From AI DisinformationOpenAI and Anthropic Pledge to S

Reading receipt

  • Apiiro reported that 1,122 public GitHub files contain secrets linked to the Don’t Ask Again feature.
  • Hacker News tracked the report with 162 points and 111 comments.
  • The source does not confirm whether Anthropic acknowledged the specific link or provided a patch.

Still open: How exactly does the Don’t Ask Again feature lead to file commits on GitHub?

Still open: Has Anthropic released a fix or configuration change to prevent this?

Time spent: 0:00 · Depth: Full

Sources

  1. Hacker News front page (trending)

Related stories

Background reading