Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessReach multipart upload endpoint
-
DeliveryCraft Content-Disposition with filename* parameter
-
ExploitValidator checks only filename, misses filename*
-
ExecutionDownstream consumes smuggled name
-
ImpactRestricted/malicious filename accepted
Vulnerability Assessment
| Exploitation Conditions | Requires an application that (1) parses multipart/form-data uploads, (2) enforces a security-relevant filename rule - such as an extension allowlist/denylist, sanitization, or path restriction - by inspecting only the standard Content-Disposition 'filename' parameter, and (3) has a downstream component that reads and acts on the RFC 2231 'filename*' extended parameter instead. The attacker supplies both parameters, placing a compliant value in 'filename' and the restricted/malicious value in 'filename*'. Limiting factors: the vulnerability is only meaningful where filename validation is used as a security control and where the 'filename*' value is actually consumed; applications that do not rely on filename rules, or that normalize both parameters consistently, are not materially affected. Exact affected product, versions, and whether the upload endpoint requires authentication are not determinable from the provided data. |
| Impact | Data for this CVE is severely limited: no CVSS score or vector, no CWE, no EPSS value, no KEV listing, and no POC indicator were supplied - the only signals are the GHSA advisory title (GHSA-5pww-8rfg-9crf) and a report via vendor:ubuntu. Based on the advisory title alone, this is a multipart filename-validation bypass: filename-restriction rules that inspect the standard Content-Disposition 'filename' parameter can be evaded by supplying an RFC 2231 extended 'filename*' parameter instead, which carries a different value that the validator does not check but a downstream consumer does. This is characteristically an integrity/validation-bypass weakness (CWE-20-class / inconsistent parsing), not a direct code-execution or memory-safety bug, so raw impact is moderate and highly dependent on how the application uses the filename (extension allowlisting, storage paths, antivirus/content checks). Exploitation status: no public exploit identified at time of analysis. Because no CVSS vector is provided, authentication requirements cannot be confirmed from the available data - this would typically be reachable by remote attackers against any endpoint accepting multipart uploads, but whether auth is required depends on the specific application. This looks like a genuine defense-in-depth-defeating parser-differential issue worth patching, but with no EPSS, KEV, or POC signal and no identified affected product, it cannot currently be ranked as a high-urgency widespread threat; priority should be set per-deployment based on whether filename rules are relied upon for a security decision. |
| Exploit ScenarioDerived | An attacker uploads a file to an endpoint that enforces filename rules (e.g. an extension denylist or sanitization) by checking the ordinary Content-Disposition 'filename' parameter. The attacker sets a benign-looking 'filename' value but smuggles a malicious name via the RFC 2231 'filename*' extended parameter, which the validator ignores while the downstream handler honors it - bypassing the restriction. No public proof-of-concept is identified in the provided data, and because no CVSS vector was supplied, the required authentication level is not confirmed. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 06, 2026
Filename-based upload restrictions can be bypassed in applications that parse multipart/form-data uploads and enforce security-relevant filename rules using only the standard Content-Disposition 'filename' parameter, because the RFC 2231 extended 'filename*' parameter can carry a different, restricted value that a downstream component may consume instead. Tracked as GHSA-5pww-8rfg-9crf and reported through the vendor:ubuntu channel, the issue carries a limited-integrity impact profile (independent assessment CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N), meaning an unauthenticated network attacker can influence which file name value is acted upon but cannot read data or cause denial of service. No affected product, version range, CPE, patch, or public exploit information was supplied, so exploitation is assessed as no public exploit identified at time of analysis.
RemediationAI
No vendor-released patch was identified at time of analysis, and no fixed version is specified in the provided data, so remediation must begin with monitoring the GHSA advisory GHSA-5pww-8rfg-9crf and the Ubuntu security notice stream for this CVE to obtain an exact patched version before upgrading. Until a fix is published, the most effective compensating control is to normalize the Content-Disposition header yourself before any validation or storage decision: parse both the legacy 'filename' token and the RFC 2231 'filename*' extended token with a single RFC 7578-compliant parser, and apply the extension allowlist, character sanitization, and path traversal checks to the resolved value that the downstream component will actually use - the trade-off is added parsing complexity and a requirement to keep the normalization layer and storage layer on the same code path, since divergence reintroduces the bug. As a stricter alternative, reject or strip any RFC 2231 extended parameters ('filename*', including continuation forms like 'filename*0*') at the edge proxy or framework boundary; this is simple and deterministic but breaks interoperability with clients that legitimately send UTF-8 or non-Latin filenames, so it should be paired with a fallback that generates a server-side filename. Additional containment: never derive storage paths directly from client-supplied filenames - generate a random server-side name and persist the original as metadata only; store uploads outside the web root and serve them through an endpoint that sets Content-Type and Content-Disposition explicitly; and, where the upload endpoint permits it, require authentication and rate-limit submissions, accepting the usability cost of authenticated uploads. Because the independent assessment rates impact as integrity-only (C:N/I:L/A:N) and no public exploit was identified, these measures reduce exposure without emergency outage-level urgency, but they should be deployed before the advisory reaches broad attention.
Affected ProductsAI
No CPE strings, product identifiers, or version ranges accompany this record, so the affected component cannot be enumerated from the available intelligence. The only provenance provided is a report via the vendor:ubuntu channel and the advisory identifier GHSA-5pww-8rfg-9crf, which indicates the disclosure is being coordinated through a GitHub Security Advisory and touches software distributed or maintained in the Ubuntu ecosystem; the GHSA URL in that advisory slug is the authoritative source for the eventual affected-package list. Because the flaw is behavioral - a mismatch between how the standard 'filename' parameter and the RFC 2231 'filename*' parameter are evaluated - any multipart/form-data parser, web framework, upload-handling library, or application that validates filenames in one layer and consumes them in another is potentially in scope, but this is an inference about the class of software rather than a confirmed affected-product statement. Consumers should treat the upstream GHSA advisory and any Ubuntu security notice (USN) issued for this identifier as the definitive source for package names and fixed versions, and should inventory their upload-processing dependencies accordingly.
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
The vulnerability lives in the parsing and resolution of the multipart/form-data Content-Disposition header, a format standardized in RFC 7578 that itself permits RFC 2231 parameter continuations and extended (asterisk-suffixed) parameters such as 'filename*' with charset and language tags. A request body can therefore legitimately contain both 'filename="photo.jpg"' and 'filename*=UTF-8''%2e%2e%2fpayload%2ephp', and different components in a request pipeline may prefer different parameters: a front-end gateway or validation middleware typically reads the plain 'filename' token to apply an extension allowlist, sanitization, or path check, while a back-end upload handler, storage layer, or framework may decode and act on the RFC 2231 'filename*' value. The result is an interpretation conflict between two parsers reading the same header, though no CWE identifier was supplied with this record so the classification is inferred from the described behavior rather than asserted by the source. The independent assessment confirms this is not a parsing crash or memory-safety issue: exploitation requires an application that (1) parses multipart uploads, (2) treats a filename rule as a security control applied only to the standard parameter, and (3) has a downstream consumer that honors 'filename*'. Where both parameters are normalized consistently, or where filename rules are not relied upon as a control, the practical impact is negligible.
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
Multipart uploads are typically network-reachable with low complexity and no user interaction; PR:N assumed absent a vector; impact is a filename-rule integrity bypass (I:L), with no inherent confidentiality or availability effect.
Estimated by vuln.today — no official severity rating has been published for this CVE yet.
DescriptionCVE.org
[GHSA-5pww-8rfg-9crf: RFC 2231 filename* parameter bypasses multipart filename rules]
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
Authentication bypass attacks exploit flaws in the verification mechanisms that control access to systems and applications.
CVE.org
GitHub PoC
Exploit-DB