mcp-tool-shop-org backpropagate up to 1.1.x backpropagate/cli.py security check

Summaryinfo

A vulnerability classified as critical was found in mcp-tool-shop-org backpropagate up to 1.1.x. This impacts an unknown function of the file backpropagate/cli.py. The manipulation results in security check. This vulnerability is known as CVE-2026-48797. It is possible to launch the attack remotely. No exploit is available. Upgrading the affected component is advised.

Detailsinfo

A vulnerability was found in mcp-tool-shop-org backpropagate up to 1.1.x. It has been classified as critical. Affected is an unknown function of the file backpropagate/cli.py. The manipulation with an unknown input leads to a security check vulnerability. CWE is classifying the issue as CWE-358. The product does not implement or incorrectly implements one or more security-relevant checks as specified by the design of a standardized algorithm, protocol, or technique. This is going to have an impact on confidentiality, integrity, and availability. CVE summarizes:

Backpropagate is a Python library for fine-tuning large language models on a single GPU. In versions 1.1.0 and 1.1.1, the optional Reflex web UI exposes a training control plane without authentication: dataset upload, model load, training start/stop, multi-run orchestration, GGUF export, and HuggingFace Hub push. The CLI accepts two operator-facing flags intended as security controls: --auth user:pass — documented as "require HTTP Basic authentication on every request to the UI." and--share — documented as "expose the UI on a public address; requires --auth." When --auth user:pass is passed, the CLI prints Auth: enabled (user: ) to confirm to the operator that authentication is active, then exports BACKPROPAGATE_UI_AUTH=user:pass to the subprocess that launches the Reflex backend. The Reflex backend (backpropagate/ui_app/**) never reads BACKPROPAGATE_UI_AUTH. No authentication middleware is registered. No request-level guard runs. No WebSocket upgrade guard runs. Any client that reaches the bound port — local or remote, depending on whether --share is used — has full UI access. An inline comment at backpropagate/cli.py:1217-1218 in the v1.1.0 source documents the gap: "For Phase 1 the variable is exported but Reflex doesn't read it yet." This comment was internal-facing; the user-facing documentation (README, CHANGELOG, SHIP_GATE) advertised the contract as enforced. An attacker who reaches the bound port can read uploaded datasets, trigger arbitrary training runs against any local base models as well as read their paths, trigger HuggingFace Hub pushes and cause disk-fill DoS. This issue has been fixed in version 1.2.0. If developers cannot immediately upgrade to 1.2.0 run backprop ui with no flags so it binds to localhost, use SSH port-forwarding (ssh -L 7860:localhost:7860 ) instead of --share for remote access, and audit any host previously launched with --share, re-issuing any HF tokens used during those sessions.

The advisory is available at github.com. This vulnerability is traded as CVE-2026-48797 since 05/22/2026. The exploitability is told to be easy. It is possible to launch the attack remotely. The exploitation doesn't require any form of authentication. Technical details are known, but there is no available exploit. This vulnerability is assigned to T1211 by the MITRE ATT&CK project.

Upgrading to version 1.2.0 eliminates this vulnerability.

If you want to get best quality of vulnerability data, you may have to visit VulDB.

Productinfo

Type

• Artificial Intelligence Software

Vendor

• mcp-tool-shop-org

Name

• backpropagate

Version

• 1.0
• 1.1

Website

• Product: https://github.com/mcp-tool-shop-org/backpropagate/

CPE 2.3info

• 🔒
• 🔒

CPE 2.2info

• 🔒
• 🔒

CVSSv4info

CVSSv3info

CVSSv2info

Exploitinginfo

Threat Intelligenceinfo

Countermeasuresinfo

Timelineinfo

Sourcesinfo

Entryinfo

If you want to get best quality of vulnerability data, you may have to visit VulDB.

Discussion