github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #15485] server/images: blob hash verification skipped when config and layer share the same digest, leads to SSRF response exfiltration #71960

New Issue
Open
opened 2026-05-05 03:09:55 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-05 03:09:55 -05:00
Owner
Copy Link

Originally created by @davidrxchester on GitHub (Apr 10, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15485

What is the issue?

When a manifest contains a config and layer with the same digest, the skipVerify map in PullModel overwrites the false value (set when the layer downloads) with true (set when the config is found already on disk). This causes verifyBlob to be skipped for both entries.

A rogue OCI registry can exploit this by serving a manifest with duplicate digests and redirecting blob downloads via 307 to internal endpoints. The SSRF response is written to disk as a blob, hash verification is skipped due to the map collision, and the blob persists. The attacker can then exfiltrate the response using /api/copy and /api/push.

This would prevent CVE-2026-5530 by preventing arbitrary 307 redirect fetches from being stored on disk.

Relevant log output

OS

macOS

GPU

Apple

CPU

No response

Ollama version

0.20.2

Originally created by @davidrxchester on GitHub (Apr 10, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15485 ### What is the issue? When a manifest contains a config and layer with the same digest, the skipVerify map in PullModel overwrites the false value (set when the layer downloads) with true (set when the config is found already on disk). This causes verifyBlob to be skipped for both entries. A rogue OCI registry can exploit this by serving a manifest with duplicate digests and redirecting blob downloads via 307 to internal endpoints. The SSRF response is written to disk as a blob, hash verification is skipped due to the map collision, and the blob persists. The attacker can then exfiltrate the response using /api/copy and /api/push. This would prevent CVE-2026-5530 by preventing arbitrary 307 redirect fetches from being stored on disk. ### Relevant log output ```shell ``` ### OS macOS ### GPU Apple ### CPU _No response_ ### Ollama version 0.20.2
GiteaMirror added the bug label 2026-05-05 03:09:55 -05:00
GiteaMirror commented 2026-05-05 03:09:57 -05:00
Author
Owner
Copy Link

@PureBlissAK commented on GitHub (Apr 18, 2026):

🤖 Automated Triage & Analysis Report

Issue: #15485
Analyzed: 2026-04-18T18:20:49.674298

Analysis

  • Type: unknown
  • Severity: medium
  • Components: unknown

Implementation Plan

  • Effort: medium
  • Steps:

This issue has been triaged and marked for implementation.

<!-- gh-comment-id:4274307247 --> @PureBlissAK commented on GitHub (Apr 18, 2026): <!-- ollama-issue-orchestrator:v1 issue:15485 --> ## 🤖 Automated Triage & Analysis Report **Issue**: #15485 **Analyzed**: 2026-04-18T18:20:49.674298 ### Analysis - **Type**: unknown - **Severity**: medium - **Components**: unknown ### Implementation Plan - **Effort**: medium - **Steps**: *This issue has been triaged and marked for implementation.*
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#71960
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?