Arbitrary File Deletion in aimhubio/aim
Description
In version 3.22.0 of aimhubio/aim, the LocalFileManager._cleanup function in the aim tracking server accepts a user-specified glob-pattern for deleting files. The function does not verify that the matched files are within the directory managed by LocalFileManager, allowing a maliciously crafted glob-pattern to lead to arbitrary file deletion.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Aim 3.22.0's LocalFileManager._cleanup accepts a user-supplied glob pattern without path validation, enabling arbitrary file deletion via crafted patterns.
Vulnerability
Overview
CVE-2024-6851 affects Aim version 3.22.0, an open-source ML experiment tracker. The LocalFileManager._cleanup function in the tracking server accepts a user-specified glob pattern to delete files but fails to verify that the matched files reside within the directory managed by LocalFileManager [2]. This allows a maliciously crafted glob pattern to escape the intended directory and delete arbitrary files on the server [3].
Exploitation
An attacker who can supply a glob pattern to the Aim server—either through an authenticated API call or an exposed endpoint—can exploit this flaw. No special privileges beyond the ability to invoke the cleanup function are required. By providing patterns such as ../../etc/passwd or /*, the attacker can delete files outside the managed directory, leading to system compromise [4].
Impact
Successful exploitation results in arbitrary file deletion on the host system. This can cause denial of service, data loss, or disruption of the Aim server and potentially other services running on the same machine. The vulnerability is particularly severe in multi-tenant or shared environments where the Aim server is accessible to untrusted users.
Mitigation
As of publication, no official patch has been confirmed for CVE-2024-6851. Users running Aim 3.22.0 should restrict network access to the Aim server, avoid exposing the cleanup endpoint to untrusted users, and monitor the project repository for updates [1]. Workarounds include disabling the cleanup functionality or implementing additional input validation at a reverse proxy.
AI Insight generated on May 20, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected packages
Versions sourced from the GitHub Security Advisory.
| Package | Affected versions | Patched versions |
|---|---|---|
aimPyPI | <= 3.22.0 | — |
Affected products
3- aimhubio/aimhubio/aimv5Range: unspecified
Patches
0No patches discovered yet.
Vulnerability mechanics
AI mechanics synthesis has not run for this CVE yet.
References
4News mentions
0No linked articles in our index yet.