CVE
Amazon DJL arbitrary file read via crafted model archive metadata
Affects Deep Java Library (DJL) from 0.7.0 through 0.38.0 when it loads model archives whose metadata names a local path. No fixed version is listed in the checked facts.
Written by anthropic/claude-haiku-5.5 from the sources below. Verify against them.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
- Exploited
- No known exploitation
- Fix
- Fix or patch referenced
- Who is exposed
- Not reported
Do first: Apply the patch listed under Evidence.
- Last modified
- Whether it needs action now or is a reference record, by fixed rules, with why. Last checked is when cve.now last re-read it.What do these mean? →
- Reference
- Last checked
- A 0-100 score for how much a CVE matters across the industry: exploitation, exposure, how widely the software is used, severity and whether a fix exists.What do these mean? →
- 28/100 · provisional
- How many real-world setups can hit the bug, from the vendor's advisory and CISA data. It does not know your setup.What do these mean? →
- Unclear
- Whether working attack code is public (Exploit-DB, Metasploit, Nuclei or GitHub). Public code makes attacks easier.What do these mean? →
- None on Exploit-DB, Metasploit or GitHub
- Exploitation: none
- Automatable: yes
- Technical impact: partial
Summary
Deep Java Library (DJL) from Amazon contains a flaw in how it handles file paths in model archive metadata. A remote attacker without a login can supply a crafted model archive to read any file the library process can access.
Original description
External control of a file path in Amazon Deep Java Library (DJL) versions 0.7.0 through 0.38.0 might allow a remote unauthenticated third party to read arbitrary files that the library process can access. To do this, the third party supplies a crafted model archive.
To remediate this issue, users should upgrade to version 0.39.0 or later.
Who should check
Teams running Deep Java Library (ai.djl:api) 0.7.0 through 0.38.0 that load model archives from outside sources, including djl-serving operators.
“Impacted versions: >=0.7.0, <=0.38.0”
cve.now Index
- Exploitation0/40No signal
- Exposure20/25No login or user action needed, CISA rates it automatable · Unknown: Advisories don't say whether default setups are affected
- Reach?/20Unknown. No tracked technology or package matched
- Severity7.5/10CVSS 7.5
- No patch0/5No signal
What would change the score
- Would score about 73 if added to CISA KEV
- Would score about 68 if CISA's SSVC rates exploitation active
- Would score about 33 if advisories confirm default setups are exposed
What to do
Apply the patch listed under Evidence.
Workaround: Only load model archives from sources you trust. Restrict network access to djl-serving management and inference APIs so untrusted archives cannot be registered.
Check your systems with an agent
Paste this into a coding or ops agent. It checks your systems read-only and reports a table per asset.
Check whether my systems are affected by CVE-2026-129266 (Arbitrary local file read via unvalidated resource location in model archive metadata in Deep Java Library), using the cve.now data below.
Rules: work read-only. Do not run exploit code, proof-of-concept scripts or scanners that send attack payloads, and do not change any configuration, package, image or deployment. Report findings only.
What cve.now has for CVE-2026-129266 (entry last updated 2026-11-16):
Affected:
- Amazon djl: from 0.7.0, through 0.38.0
Fixed versions: none listed yet.
Conditions (all must be true for the bug to be hit):
- Affected DJL versions from 0.7.0 through 0.38.0 per the AWS bulletin (the GitHub advisory lists <= 0.39.0). (source: https://aws.amazon.com/security/security-bulletins/2026-183-aws/)
- DJL loads a model archive whose metadata names a local path as a resource location. (source: https://github.com/deepjavalibrary/djl/security/advisories/GHSA-xf3w-qp3m-fp62)
Setup: unknown whether the bug needs a non-default feature or setting. Check the advisory.
Attacker needs: network access, no login, no user action.
Workaround: Only load model archives from sources you trust. Restrict network access to djl-serving management and inference APIs so untrusted archives cannot be registered. (source: https://aws.amazon.com/security/security-bulletins/2026-183-aws/)
CISA KEV: CVE-2026-129266 is not listed as of cve.now's last read.
This KEV status is for CVE-2026-129266 only. If you also check another CVE, look up its own KEV status; never assume it from this one.
How to check:
- Installed software: check what is actually running, not what a repo declares. Read the version from each host, container image and Kubernetes node (the binary's version output, the OS package database, image layers), and check whether the conditions above hold there. Distribution packages may carry a backported fix under an older version number; confirm against the distribution's own advisory.
Verify against the linked advisories before concluding, and cite the date you read each one:
- https://aws.amazon.com/security/security-bulletins/2026-183-aws/
- https://github.com/deepjavalibrary/djl/releases/tag/v0.39.0
- https://github.com/deepjavalibrary/djl/security/advisories/GHSA-xf3w-qp3m-fp62
Do not conclude an asset is unaffected unless you checked it and the evidence covers both its version and the conditions above. If you could not check something, mark it unknown and say what is missing.
Report a table with one row per asset and these columns: Asset | Version or build | Status (affected, fixed, not applicable, unknown) | Compensating controls | Exposure (internet-facing, internal, none) | Evidence | Next action.
Entry: https://cve.now/CVE-2026-129266
For follow-up questions, use the cve.now MCP server at https://mcp.cve.now/mcp (no authentication), for example execute with name cve.get and arguments {"id": "CVE-2026-129266"}.Timeline
- CVE published
- First seen by cve.now
- CISA SSVC assessment
Exploitation: none · Automatable: yes · Technical impact: partial
- Record last updated
Affected versions
- Amazon djl
From 0.7.0 · through 0.38.0
Evidence 5
Vendor advisory · 2
Establishes: Affected and fixed versions.
- Advisory
- Advisory
CVE record · 2
Establishes: The identifier and description.
- Record
- Record
Other links · 1
- Patch
Corrections
None so far.
Something wrong on this page? Report an error. Every fix is listed here.
Questions about CVE-2026-129266
Is CVE-2026-129266 being exploited?
It is not in CISA's Known Exploited Vulnerabilities catalog, and CISA's SSVC assessment does not mark exploitation as active. That does not mean it has not been attacked: vendors and researchers can report attacks before CISA lists them. No public exploit code is tracked.Source: cve.now feed data
What does this mean?
DJL does not validate the scheme of a resource location in model archive metadata. A crafted archive that names a local path makes DJL open and read that file, returning its contents as model data to the caller. The file read can reach any file the DJL library process can access.Source: aws.amazon.com, github.com
Which versions are affected by CVE-2026-129266?
djl: from 0.7.0 through 0.38.0.Source: cve.now feed data
Am I affected, and how do I check?
The AWS bulletin lists DJL versions 0.7.0 through 0.38.0; the GitHub advisory lists versions up to and including 0.39.0. The affected setup is loading model archives whose metadata names a local path as a resource location. Check the djl version in the project's dependency configuration to see which version is in use.Source: aws.amazon.com, github.com
How do I fix CVE-2026-129266?
The checked facts name no fixed version and no patch ids. The vendor's workaround is to load model archives only from sources you trust, and to restrict network access to the djl-serving management and inference APIs so untrusted archives cannot be registered.Source: aws.amazon.com, github.com