Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessPublish malicious git-annex repo with attacker remote
-
DeliveryVictim clones repo
-
Exploitgit-annex autoenables attacker-controlled special/git remote
-
ExecutionVictim runs git-annex sync --content
-
ImpactPrivate annexed files exfiltrated to attacker
Vulnerability Assessment
| Exploitation Conditions | Exploitation requires three concrete preconditions stated in the advisory: (1) the attacker must control one of the victim's git-annex remotes, typically by supplying a git-annex repository that the victim clones; (2) the attacker must already have obtained information about the victim's private repository - either specific annexed keys/files known to be present, or the network location of the victim's private git repository; and (3) the victim must run a content-transferring command such as |
| Impact | No CVSS vector, EPSS score, KEV listing, or SSVC technical-impact decision is provided; the vuln.today priority is explicitly 'awaiting_data', so quantitative signals are absent and the following is derived from the vendor advisory text alone. The flaw is a genuine confidentiality breach - cross-repository private data exfiltration - but exploitation is conditional rather than trivial: the attacker must already control one of the victim's git-annex remotes (e.g., supply a repository the victim clones) AND must already possess specific knowledge of the victim's private repository (which annexed keys/files exist, or the location of the private git repo), AND the victim must enable the remote and run a content-transferring command such as |
| Exploit ScenarioDerived | An attacker publishes a malicious git-annex repository and waits for a victim to clone it; on initialization git-annex autoenables an attacker-configured special remote (or git remote) that points at the victim's own private data store. When the victim later runs |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 26, 2026
Private annexed content can be exfiltrated from a victim's git-annex repository through attacker-controlled remotes, allowing data the user believes is private (annexed files kept out of a published git repo, or objects in a private S3 bucket or other special remote) to be pulled and sent to the attacker. The assessed vector is CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:N/A:N: remote, unauthenticated in the sense that no credentials to the victim's systems are needed, but requiring user interaction and three preconditions from the vendor advisory - the attacker must control one of the victim's remotes, must already possess reconnaissance about the private repository (specific annexed keys, or the location of the private git repository), and the victim must run a content-transferring command such as git-annex sync --content against a special remote autoenabled on init or manually enabled and then retargeted. Versions before 10.20261026 are affected; no CISA KEV listing, no public exploit code, and no EPSS score were available at time of analysis, and the vuln.today priority label is 'awaiting_data', so exploitation has not been observed.
RemediationAI
Vendor-released patch: upgrade to git-annex 10.20261026 from git-annex.branchable.com/download/, or apply the backported branch matching your release line (CVE-2026-112818-patch-10.20250416, CVE-2026-112818-patch-10.20251029, or CVE-2026-112818-patch-10.20261006) from the git-annex git repository. Updating is the complete fix as stated in the vendor advisory; the patched code makes git-annex enableremote print the special remote's configuration for the user to verify before the remote can be used, and restricts how autoenabled special remotes may be used, so after upgrading, expect an interactive confirmation step the first time a special remote is enabled and treat any unexpected prompt as a signal that the repository configuration deserves inspection. If upgrading cannot be done immediately, the compensating control that follows directly from the advisory is to avoid initializing git-annex in or running content-transferring commands such as git-annex sync --content inside repositories cloned from untrusted sources until patched, since all three attack variants depend on that workflow; the trade-off is that this blocks legitimate use of third-party git-annex repositories, and it does not protect users who manually enable a special remote later retargeted by the attacker. Reference URLs: https://seclists.org/oss-sec/2026/q4/394 and https://git-annex.branchable.com/security/cross_repository_data_exfiltration_via_special_remote/.
Affected ProductsAI
git-annex releases prior to 10.20261026 are affected, per the oss-security advisory and the vendor security page at git-annex.branchable.com/security/cross_repository_data_exfiltration_via_special_remote/. The maintainer published fixes on the current release line and as backport branches in the git-annex git repository for the 10.20250416, 10.20251029 and 10.20261006 releases (CVE-2026-112818-patch-10.20250416, CVE-2026-112818-patch-10.20251029, CVE-2026-112818-patch-10.20261006), so installations tracking older distribution or cabal/stack packages should map their version against those branches rather than assuming only the newest line was patched. No CPE dictionary entries were supplied, so affected installations must be identified by git-annex version and by whether they clone untrusted git-annex repositories or enable third-party special remotes. Download locations are listed at git-annex.branchable.com/download/.
Affects your stack
Find out whether the stacks and servers you run are on an affected version.
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
git-annex manages file content outside of git itself, storing objects under content-hash 'keys' and tracking their locations in a separate annex; remotes - including 'special remotes' backed by S3 buckets, rsync, web or external helper programs - describe where those objects can be fetched from, and a repository's configuration can carry special-remote definitions that are autoenabled when git-annex is initialized in a clone. The flaw is in that trust model rather than in a memory-safety primitive: a cloned repository is configuration-controlled input, so an attacker who supplies the repository can supply remote definitions and a tree referencing keys, and once the victim asks git-annex to transfer content, git-annex will fetch objects from whichever remote it has been told about. The advisory documents three variants - autoenabled special remote pointed at the victim's private annex, a benign-looking special remote whose configuration is later retargeted at the victim's own bucket, and an autoenabled git remote pointing at the victim's private repository that discloses key names on the first sync and exfiltrates objects on the second. No CWE was assigned in the supplied data and no CVSS vector was published by the vendor; the assessed vector above reflects analyst judgment about the reconnaissance and interaction prerequisites. The upstream fix hardens the configuration-trust boundary: git-annex enableremote now displays the special remote's configuration and requires the user to verify it is safe before use, and the use of autoenabled special remotes is restricted.
Know when the next one hits your stack
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
Attacker supplies a remote remotely (AV:N) but needs prior knowledge of the victim's private data and victim-run sync (AC:H, UI:R); crosses into a separate private store (S:C) with confidentiality-only impact (C:H, I/A:N).
Estimated by vuln.today — no official severity rating has been published for this CVE yet.
Description PRE-NVD
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
Attack Technique
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
CVE.org
GitHub PoC
Exploit-DB