INSPECTING ARCHIVED INTELLIGENCE (OUTDATED VERSION).
AI Coding Agents Leaked Secrets to GitHub via Supply Chain
|
2026-11-10 18:46
HIGH
HIGH
DATA BREACH
Executive Summary
AI-generated
On November 10, 2026, thousands of secrets were leaked to GitHub public repositories due to a vulnerability in AI coding assistants like Claude Code. Researchers from security firm Apiiro found 2,539 credentials across nearly 2,000 public GitHub repositories, including database passwords, cloud keys, API tokens, and SSH root logins. The attack works by using the "Yes, and don't ask again" prompt to save secrets to a hidden file (.claude/[IOC HIDDEN • LOGIN REQUIRED]) inside the project folder, which is then committed with the rest of the code. As a result, 1,122 files contained real secrets, including 636 provider-specific keys and 743 passwords or tokens associated with public hosts.
Technical Mitigations AI-generated
• Reset credentials and API keys that may have been exposed, and review access logs for the affected accounts or integrations.
Technical Observables 1 IoCs Login Required
Indicators of Compromise (IoCs) Protected
Raw threat telemetry, malicious IP addresses, file hashes, and direct VirusTotal correlation are restricted to authenticated users.
files 1
Intelligence Metadata
Actors / Malware / CVEs / Campaigns
Target & Sectors
Global Scope
technologytechnology
Incident Timeline
January 2025
Nearly 1.5 million exposed secrets were targeted in a revocation program that revealed the existence of previously non-existent working keys in AI coding tools' config files since January 2025.
Click on any entity below to view its context and source!
general_metric
2,655 working keys
“Since January 2025 we’ve found 2,655 working keys in AI coding tools’ config files like that one,” the researchers said, noting that this was a non-existent problem the year before.
2025/11/05
“Telling developers their keys leaked doesn’t get them turned off: 81% of the keys we reported last year still worked six months later,” the company explained .
Click on any entity below to view its context and source!
general_metric
81 %
“Telling developers their keys leaked doesn’t get them turned off: 81% of the keys we reported last year still worked six months later,” the
company explained
.
2026/11/10
In 539 of the 1,105 repositories containing real secrets that the researchers investigated, a commit co-authored by Claude Code had modified the affected settings file.
Click on any entity below to view its context and source!
organisation
Adaptive Security's
Adaptive Security's AI agents analyze email intent, context, and behavior to uncover phishing, business email compromise, and invoice fraud.
data_breach
1,122 files
In the case of Claude Code, Apiiro examined 104,185 publicly accessible copies of .claude/settings.local.json on GitHub, finding that 1,122 files contained real secrets.
organisation
AI Coding Agents Leak Thousands
AI Coding Agents Leak Thousands of Secrets to GitHub.
organisation
Apiiro
Apiiro tracked some of the leaked data to
Claude Code
, which was, unbeknownst to users, copying secrets to a hidden file and committing them with the rest of the developer's code.
organisation
Shmulik Cohen
Among specific secrets leaked were "a database vendor's own admin-level cloud key, a production Azure Postgres password, a Square production token, and the owner password of a medical-records database," Apiiro researchers Shmulik Cohen and Matan Giladi wrote in the post.
organisation
Matan Giladi
Among specific secrets leaked were "a database vendor's own admin-level cloud key, a production Azure Postgres password, a Square production token, and the owner password of a medical-records database," Apiiro researchers Shmulik Cohen and Matan Giladi wrote in the post.
organisation
BreachLock
The action that exposes these secrets while using Claude Code appears to be a "dead zone" of sorts for developers using AI coding assistants, so they likely have no idea that they've just exposed a secret, observes Seemant Sehgal, founder and CEO of AI penetration testing firm BreachLock.
data_breach
2,539 credentials
Research from security firm Apiiro found
2,539 credentials
across nearly 2,000 public GitHub repositories, with secrets such as database passwords, cloud keys, API tokens, and SSH root logins among the exposed data, according to
a recent report
published online.
victims
63 organizations
People Are Getting Weird About AI
Most of the secrets belong to individual developers across 63 organizations, "from student groups to startups, software agencies, and established vendors," they wrote.
organisation
Microsoft 365
See what's getting through with a 14-day read-only scan of Microsoft 365 or Google Workspace, with no MX changes.
organisation
Google Workspace
See what's getting through with a 14-day read-only scan of Microsoft 365 or Google Workspace, with no MX changes.
organisation
Truffle Security
Truffle Security has launched the Secrets Revocation Program, a free initiative that helps API key issuers revoke 1.4 million publicly exposed secrets that remain valid.
organisation
the Secrets Revocation Program
Truffle Security has launched the Secrets Revocation Program, a free initiative that helps API key issuers revoke 1.4 million publicly exposed secrets that remain valid.
organisation
OpenAI’s July
AI is also increasing the impact of exposure on the offensive side, as shown by OpenAI’s July incident, where AI agents used
exposed Hugging Face tokens
to steal data from the platform.
organisation
Truffle
Revoking exposed keys
Truffle’s new program, which uses the TruffleHog tool to check for the validity of leaked keys, aims to coordinate the revocation of exposed secrets with the entities that issued them.
organisation
TruffleHog
Revoking exposed keys
Truffle’s new program, which uses the TruffleHog tool to check for the validity of leaked keys, aims to coordinate the revocation of exposed secrets with the entities that issued them.
organisation
Hugging Face, Notion
Key removal procedure
Source: Truffle Security
Sixteen issuers have already joined the program, including Hugging Face, Notion, Vercel, Resend, RubyGems, and Twilio.
organisation
RubyGems
Key removal procedure
Source: Truffle Security
Sixteen issuers have already joined the program, including Hugging Face, Notion, Vercel, Resend, RubyGems, and Twilio.
organisation
Twilio
Key removal procedure
Source: Truffle Security
Sixteen issuers have already joined the program, including Hugging Face, Notion, Vercel, Resend, RubyGems, and Twilio.
organisation
Truffle Security's
Companies that issue API keys can
join Truffle Security's Secrets Revocation Program
to automatically protect their customers.
organisation
Secrets Revocation Program
Companies that issue API keys can
join Truffle Security's Secrets Revocation Program
to automatically protect their customers.
data_breach
543,000 valid credentials
Earlier this year, Truffle Security reported that it had found
543,000 valid credentials
leaked in public GitHub repositories.
Tactical Metrics
Metrics
data_breach
1,122
Files
Click for context!
In the case of Claude Code, Apiiro examined 104,185 publicly accessible copies of .claude/settings.local.json on GitHub, finding that 1,122 files contained real secrets.
Metrics
data_breach
2,539
Credentials
Research from security firm Apiiro found
2,539 credentials
across nearly 2,000 public GitHub repositories, with secrets such as database passwords, cloud keys, API tokens, and SSH root logins among the exposed data, according to
a recent report
published online.
Metrics
victims
63
Organizations
People Are Getting Weird About AI
Most of the secrets belong to individual developers across 63 organizations, "from student groups to startups, software agencies, and established vendors," they wrote.
Metrics
infrastructure
Microsoft 365
Affected Product
See what's getting through with a 14-day read-only scan of Microsoft 365 or Google Workspace, with no MX changes.
Metrics
data_breach
543,000
Valid Credentials
Earlier this year, Truffle Security reported that it had found
543,000 valid credentials
leaked in public GitHub repositories.
Intelligence Sources
BleepingComputer
2026-11-05
Dark Reading
2026-11-10
AI Coding Agents Leak Thousands of Secrets to GitHub
Dark Reading
Unpublish from Social Media?
Are you sure you want to delete this podcast video from all synchronized social networks (YouTube, Facebook, Threads)?
Important:
Due to Meta API restrictions, Instagram Reels cannot be deleted automatically via API by third-party apps.
View Profile to Delete Manually
View Profile to Delete Manually
Tactical Intelligence
Report Intelligence Issue
Podcast Options
Generate
Incident Version History
CURRENT VERSION
Last Updated: 2026-11-11T07:25
Comprehensive Tactical Telemetry
Highly Correlated Entities
19x
organisation
Identified Entity
AI Coding Agents Leak Thousands
entity
5x
timeline
Temporal Reference
June 16, 2024
date
3x
general metric
%
81
%
2x
tactic
Cyber Operation Type
Data Leak
tactic
Contextual Telemetry
Context Block
16 METRICS
industry
Targeted Sector
Technology
sector
tactic
MITRE ATT&CK Technique
T1566 - Phishing
technique
general metric
Accessible Copies
104,185
accessible copies
data breach
Files
1,122
files
data breach
Credentials
2,539
credentials
general metric
Public Github Repositories
2,000
public github repositories
victims
Organizations
63
organizations
general metric
Distinct Secrets
1,379
distinct secrets
general metric
Specific Keys
636
specific keys
general metric
Passwords
743
passwords
general metric
Repositories
1,105
repositories
infrastructure
Affected Product
Microsoft 365
software
general metric
Exposed Secrets
1,400,000
exposed secrets
general metric
Working Keys
2,655
working keys
data breach
Valid Credentials
543,000
valid credentials
general metric
Aws Secrets
431,000
aws secrets
Click on any entity below to view its context in the main text!
Selective Unpublish
Select the networks you want to remove this post from. The system will try to delete the real post through the API and clean the database so you can publish it again.
By navigating this website, you accept the use of strictly necessary technical cookies for session security and basic platform functionality. We do not use tracking or advertising cookies. Read our Privacy Policy.
Generating report…
0s