TL;DR
Get business pricing on networking and server gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Apiiro reports that passwords appeared in 1,122 publicly accessible GitHub files associated with Claude Code’s “Don’t Ask Again” feature. The supplied report does not identify the affected accounts, explain how the files were collected, or say whether the credentials remain active or have been revoked.
Security company Apiiro says it found passwords in 1,122 publicly accessible GitHub files associated with Claude Code’s “Don’t Ask Again” behavior, raising concerns about sensitive information entering repositories through AI-assisted coding workflows. The supplied report extract does not specify when the files were identified or whether Anthropic, GitHub, or affected account holders have been notified.
The reported total is 1,122 public files, not necessarily 1,122 distinct accounts or unique passwords. The material provided from Apiiro’s report does not describe the files, establish how many contained valid credentials, or state whether the passwords provided access to production systems, developer accounts, or other services. The number should be read as Apiiro’s finding, not as an independently verified count.
The report’s title links the alleged exposure to Claude Code’s “Don’t Ask Again” option. The source extract does not explain the feature’s precise operation in the cases studied, show how passwords were written to files, or document a reproducible sequence leading to publication. It also does not include comments from Anthropic or GitHub. Those missing details limit what can be concluded about the tool’s role and the circumstances of the exposures.
Apiiro’s supplied page text identifies the company as an AI application-security vendor, but the provided material contains no methodology, sample files, scan dates, credential-validation results, or breakdown by repository. Without those details, readers cannot assess the detection scope, the share of findings that were live secrets, or whether the files have since been removed.
Risks for Exposed Credentials
Passwords published in a public repository can be copied and retained by others, even if a file is later deleted. If the credentials are active and grant access to a service, repository, or account, the owner may need to revoke them and check for unauthorized use. The source does not establish that any of the reported passwords remain valid or were used by anyone.
The finding also points to a broader operational concern: developers may use coding agents in environments where project files, command output, or saved settings contain sensitive material. The report’s headline associates the issue with a Claude Code option, but the available evidence does not establish that the software itself disclosed passwords or that the behavior affects all users. The distinction matters when assessing both user risk and any responsibility for preventing secrets from reaching public code.
For organizations, the reported count is a reason to review repository controls and secret-scanning practices, not proof of a confirmed breach. A public file containing a password can create exposure, but determining impact requires knowing whether the credential worked, what it could access, and whether it was used.
Products Worth Considering
Samsung Galaxy Tab A11+ 8GB RAM, 256GB Storage, Gray
Samsung Galaxy Tab A11+ 11-Inch Tablet, 6GB RAM, 128GB Storage, Expandable, Gray
Bose Lifestyle Ultra Compact Wireless Smart Home Speaker Bundle, Black
As an affiliate, we earn on qualifying purchases.
The Report’s Available Evidence
The source is an Apiiro report titled “Coding Agents Leaking Secrets: GitHub ‘Don’t Ask Again.’” The supplied text includes extensive website navigation and promotional descriptions of Apiiro’s security products, but little of the report’s underlying evidence. It does not provide a timeline, explain how the files were located, or identify particular repositories or affected services.
That gap means the reported count cannot be independently evaluated from the supplied material. In particular, a file containing a string that resembles a password is not by itself proof that the credential is genuine, current, or exploitable. The extract also does not say whether the files were created by Claude Code, merely contained references to its feature, or were connected to it in another way.
The distinction is important because the headline combines a specific tool behavior with a precise number. Apiiro is the source for both claims here; no separate confirmation, response from Anthropic, or GitHub statement is included in the material available for this article.
““Claude Code’s ‘Don’t Ask Again’ exposes passwords in 1,122 public GitHub files.””
— Apiiro report headline
Questions About the Files
Several central details remain unknown: when Apiiro scanned the repositories, how it defined a password, whether credentials were tested, and whether the 1,122 files represent separate repositories or repeated copies. The source extract does not disclose the files or provide a technical account of how the “Don’t Ask Again” behavior was connected to their contents.
It is also unclear whether any secrets were still active when found, what systems they could reach, whether anyone accessed them, or whether repository owners were contacted. The material does not say if the files have been removed, if credentials were rotated, or if Anthropic or GitHub has investigated the report. No response from either company is included.
These gaps prevent a firm assessment of the scale of any practical compromise. The confirmed point in the supplied material is that Apiiro made the allegation and gave a count of 1,122 files; the reported credential validity, cause, and impact have not been established here.
Verification and Credential Response
The next useful information would be Apiiro’s full methodology, scan dates, examples with sensitive values safely redacted, and a breakdown showing how it verified the files and credentials. Responses from Anthropic and GitHub could clarify whether the feature was involved, what safeguards or investigations are in place, and whether affected repository owners have been notified.
Owners of repositories identified in a fuller report would need to determine whether exposed credentials are genuine, revoke or rotate any active ones, and review available access records. Removing a file alone may not address exposure if its contents were copied or retained in repository history. The supplied material does not confirm that any such response has occurred.
Until those details are available, the reported figure should be treated as an attributed security finding rather than a confirmed count of compromised accounts. Further reporting should distinguish detected password-like strings from validated, active credentials and document any confirmed misuse.
Key Questions
What did Apiiro report?
Apiiro said passwords appeared in 1,122 publicly accessible GitHub files associated with Claude Code’s “Don’t Ask Again” behavior. The supplied material does not independently verify the count.
Does this mean 1,122 accounts were compromised?
No. The figure refers to files, not confirmed accounts, unique credentials, or successful account access. The source extract does not say how many passwords were valid or whether anyone used them.
How were passwords exposed?
The available report material does not describe the specific mechanism or provide evidence showing how the feature led to the files. The connection is presented in Apiiro’s headline, but the details remain unclear.
Have Anthropic or GitHub responded?
No response from Anthropic or GitHub appears in the supplied source material. It also does not say whether either company was notified.
What should repository owners do if they find a password?
They should treat a genuine exposed credential as potentially compromised: revoke or rotate it, check what it could access, and review relevant account or service activity. Deleting the visible file may not remove copies from repository history or from others’ systems.
Source: apiiro.com
