TorchServe is a tool for serving and scaling PyTorch models in production. TorchServe default configuration lacks proper input validation, enabling third parties to invoke remote HTTP download requests and write files to the disk. This issue could be taken advantage of to compromise the integrity of the system and sensitive data. This issue is present in versions 0.1.0 to 0.8.1. A user is able to load the model of their choice from any URL that they would like to use. The user of TorchServe is responsible for configuring both the allowed_urls and specifying the model URL to be used. A pull request to warn the user when the default value for allowed_urls is used has been merged in PR #2534. TorchServe release 0.8.2 includes this change. Users are advised to upgrade. There are no known workarounds for this issue.
CVE-2023-43654
“—” beside the score: it is equal to NVD's CVSS score.
This critical-severity CVE scores 9.8 under NVD CVSS v3. EPSS exploit probability: 42% (98th percentile of EPSS-scored CVEs). No GitHub Security Advisory CVSS is on record for this CVE.
- Public exploit code is available (Metasploit, other public proof-of-concept code)
A fix is available — apply it.
- CVSS v3
- 9.8
- EG Score
- 9.8CRITICALmedium confidence
- EG Risk
- 85EG Risk 85/100
EG Risk is EchelonGraph's 0–100 priority score. It fuses impact severity (EG Severity, the CVSS base score, never the KEV-floored or EPSS-lifted EG Score) with real-world exploitation and automatability, under weights EchelonGraph set by hand, so you can rank equal-severity CVEs and fix the most dangerous first. Higher = act sooner.
How it’s computed (value × EchelonGraph’s weight)Severity98% × 45%Impact severity (EG Severity): CVSS 9.8 of 10. Never the KEV-floored or EPSS-lifted EG Score.Exploitation65% × 40%EchelonGraph's fixed weight for EG-KEV Tier 2 (Weaponized): a constant, not a probability.Automatability100% × 15%CISA's SSVC Automatable value: yes (Vulnrichment).SSVC (CISA's decision table, applied by EchelonGraph): Track at low, Track* at medium and Attend at high mission impact.Exploitation public PoC (CISA Vulnrichment) · Automatable yes (CISA Vulnrichment) · Technical impact total (CISA Vulnrichment). All three are CISA's SSVC values (Vulnrichment). EchelonGraph applied CISA's decision table to them; CISA publishes SSVC inputs and the table, not a decision for each CVE.Action: A fix is available. Apply it within your standard update timelines at low or medium mission impact and sooner than that at high. - EPSS PROB
- 42%
- EPSS %ILE
- 98th
- KEV
- Not listed
What the EG score is made of (scorer v4.0: the three questions the 0–10 score mixes, published apart · methodology)
The EG score (9.8): severity, escalated by confirmed exploitation. This CVE's 9.8 is its impact severity (EG Severity), not escalated.
- EG Severity
- 9.8 CRITICALCVSS-B, impact only: no KEV listing, EPSS or EG-KEV tier raises it.NVD CVSS v3 · CVSS v3.0 or v3.1
- EG Likelihood
- 42% within 30 daysSource: FIRST EPSS (model v2026.06.15, score date 2026-10-22). FIRST's exploitation probability, not an EchelonGraph model.
- Confirmed exploitation
- None on record (CISA KEV, EG-KEV Tier 1)Reported apart from the likelihood: an observed fact, not a probability.
- EG Priority
- SSVC Track* at medium mission impactTrack at low, Track* at medium and Attend at high mission impact.CISA BOD 26-04: 3 days if the affected system is publicly exposed; 60 days if it is not.Each input and its source: the SSVC and BOD 26-04 blocks below.
SSVC (CISA's decision table, applied by EchelonGraph)Track at low, Track* at medium and Attend at high mission impact.
A fix is available. Apply it within your standard update timelines at low or medium mission impact and sooner than that at high.
Exploitation public PoC (CISA Vulnrichment) · Automatable yes (CISA Vulnrichment) · Technical impact total (CISA Vulnrichment). All three are CISA's SSVC values (Vulnrichment). EchelonGraph applied CISA's decision table to them; CISA publishes SSVC inputs and the table, not a decision for each CVE. Mission impact is CISA's Mission & Well-being decision point, and only you can judge it. CISA's table makes it high when Mission Prevalence is Essential — the vulnerable component "directly provides capabilities that constitute at least one MEF for at least one entity" (MEF: mission essential function) — or when Public Well-Being Impact is Irreversible: "multiple fatalities are likely", the cyber-physical system "is likely lost or destroyed", "extreme or serious externalities" are imposed on other parties, or social systems such as elections or the financial grid "are destabilized and potentially collapse". CISA's decision table
BOD 26-04 (CISA's remediation timeline, Table 1)3 days if the affected system is publicly exposed; 60 days if it is not.
In CISA KEV no (CISA's KEV catalog) · Automatable yes (CISA Vulnrichment) · Technical impact total (CISA Vulnrichment). Whether your system is publicly exposed is yours to answer. BOD 26-04 binds US Federal Civilian Executive Branch agencies; for anyone else it is CISA's published timeline, a reference rather than an obligation. The directive · how EchelonGraph applies it
Published
September 28, 2023
Last Modified
June 17, 2026
Advisory Details (4)
Auto-updated Oct 17, 2026TorchServe SSRF · Advisory · pytorch/serve · GitHub
https://github.com/pytorch/serve/security/advisories/GHSA-8fxr-qfr9-p34wTorchServe v0.8.2 Release Notes
Patch available: pytorch/serve v0.8.2
https://github.com/pytorch/serve/releases/tag/v0.8.2Issue warning about allowed_urls when default value is used
Fix merged in pytorch/serve PR #2534 on 2023-08-22 — awaiting tagged release
https://github.com/pytorch/serve/pull/2534Affected Packages
(1 across 1 ecosystem)
PyPI(1)
| Package | Vulnerable range | Fix by version range | Dependents |
|---|---|---|---|
| torchserve | 0.1.1 ... 0.8.1 (18 versions) |
| — |
Weakness Classification(1)
MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.
Data Freshness Timeline
(refreshed 13× in last 7d / 36× in last 30d)
Each row is a time one of our pipelines fetched, updated or re-scored this CVE. For example, "NVD update" means NVD published or revised its analysis for this CVE; "MITRE cvelistV5" means we ingested or refreshed it from the CNA feed. An arrow marks a change the record states ("CVSS v3 → 9.1"); EPSS and GHSA rows show the values recorded at that time, not what changed (EPSS values are rounded; hover one for the exact figures). Most recent first.
Showing the most recent 86 of 154 refreshes on record for this CVE.
A repeat, a row identical to one the same pipeline wrote less than 10 seconds earlier, is not a separate refresh and is not counted. 14 repeat rows among the latest 100 are merged into the row they repeat. The 68 oldest rows on record are not shown and were not checked, so the counts can still include repeats.
- 2026-10-23 16:29 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-23 06:25 UTCEG score recompute
- 2026-10-22 16:31 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-22 05:47 UTCEG score recompute
- 2026-10-21 16:29 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-21 05:06 UTCEG score recompute
- 2026-10-20 16:30 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-20 04:28 UTCEG score recompute
- 2026-10-19 16:29 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-19 03:50 UTCEG score recompute
- 2026-10-18 03:12 UTCEG score recompute
- 2026-10-17 16:30 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-17 02:34 UTCEG score recompute
- 2026-10-16 16:36 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-16 01:55 UTCEG score recompute
- 2026-10-15 16:29 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-14 21:23 UTCEG score recompute
- 2026-10-14 16:30 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-13 17:24 UTCEG score recompute
- 2026-10-13 16:30 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-12 16:44 UTCEG score recompute
- 2026-10-11 21:45 UTCEG score recompute
- 2026-10-11 13:45 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-09 20:05 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-08 17:06 UTCEPSS rescoreEPSS 42% · 98th percentile
Show 61 moreShow fewer
- 2026-10-07 23:46 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-06 15:15 UTCEPSS rescoreEPSS 44% · 98th percentile
- 2026-10-03 14:23 UTCEPSS rescoreEPSS 42% · 98th percentile
- 2026-10-01 19:48 UTCEPSS rescoreEPSS 41% · 98th percentile
- 2026-09-30 16:31 UTCEG score recompute
- 2026-09-30 15:01 UTCEPSS rescoreEPSS 41% · 98th percentile
- 2026-09-29 15:17 UTCEPSS rescoreEPSS 40% · 98th percentile
- 2026-09-27 13:42 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-26 15:56 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-26 00:22 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-24 14:01 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-23 13:46 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-22 22:01 UTCOSV refresh
- 2026-09-18 19:25 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-17 19:25 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-16 14:06 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-16 05:11 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-12 14:58 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-11 14:50 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-11 09:34 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-08 21:58 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-06 13:45 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-05 15:26 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-04 13:58 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-04 05:04 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-03 17:16 UTCOSV refresh
- 2026-09-02 14:09 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-09-01 13:51 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-30 19:15 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-30 01:20 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-28 21:39 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-26 14:43 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-25 13:46 UTCEPSS rescoreEPSS 36% · 98th percentile
- 2026-08-21 23:47 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-20 22:53 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-19 17:02 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-18 13:46 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-17 06:27 UTCOSV refresh
- 2026-08-16 14:54 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-16 02:12 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-15 01:28 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-13 21:58 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-12 13:48 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-11 13:39 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-07 16:25 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-06 13:45 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-05 19:15 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-04 15:07 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-08-02 01:06 UTCOSV refresh
- 2026-08-01 04:13 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-30 16:26 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-28 15:34 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-26 14:52 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-25 14:16 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-24 14:16 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-23 14:16 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-23 02:26 UTCEG score recompute
- 2026-07-22 23:13 UTCEG score recompute▼ 0.20
- 2026-07-22 14:06 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-21 15:23 UTCEPSS rescoreEPSS 35% · 98th percentile
- 2026-07-20 17:06 UTCEPSS rescoreEPSS 35% · 98th percentile
Publicly available exploits
(4 references)Public exploit code is referenced for this CVE (1 Metasploit module, 2 GitHub PoCs), as is 1 Nuclei detection template. A detection template checks whether a system is affected; it is not necessarily exploit code. Defenders should treat patch urgency accordingly.
- Metasploitexploit/multi/http/torchserver_cve_2023_43654✓ verifiedPublished Oct 12, 2023
PyTorch Model Server Registration and Deserialization RCE
Open source ↗ - Nucleihttp/cves/2023/CVE-2023-43654.yamlPublished Sep 19, 2024
PyTorch TorchServe SSRF
Open source ↗ - GitHub PoCOligoCyberSecurity/CVE-2023-43654Published May 28, 2024
Demo for CVE-2023-43654 - Remote Code Execution in PyTorch TorchServe
Open source ↗ - GitHub PoCOligoCyberSecurity/ShellTorchCheckerPublished Oct 2, 2023
A tool that checks if a TorchServe instance is vulnerable to CVE-2023-43654
Open source ↗
Related CVEs(same product + same CWE)
Same product
4 shownPyPI:torchserve
Same CWE
10 shownCWE-918
Frequently asked(5)
What is CVE-2023-43654?
When was the CVE record for CVE-2023-43654 published?
Is CVE-2023-43654 actively exploited?
What is the CVSS score of CVE-2023-43654?
How do I remediate CVE-2023-43654?
Dependency Blast Radius
See which npm, PyPI, Go, and Maven packages are affected by CVE-2023-43654
Is Your Infrastructure Affected by CVE-2023-43654?
EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.