py-libp2p's yamux implementation contains a remote denial-of-service vulnerability that warrants your immediate attention. The issue is that py-libp2p reads the declared body length from a yamux frame header, then performs a blocking read for that exact number of bytes — and only AFTER the body arrives does it check whether the length exceeds MAX_WINDOW_SIZE. A malicious peer can send a frame with an arbitrarily large declared length, causing your process to block indefinitely waiting for data that will never arrive, or until TCP timeouts eventually free the connection. The attack works on the default multiplexer, meaning one 12-byte frame from an authenticated peer freezes every stream multiplexed over that connection simultaneously. The root cause is architectural: yamux was designed in Go with the assumption that blocking I/O is acceptable because goroutines are cheap. py-libp2p inherited this pattern without adapting it for Python's async context, creating a situation where a length field intended for flow control is being used as an implicit security boundary — which it isn't. The yamux spec treats MAX_WINDOW_SIZE as a buffer limit, not a DoS prevention mechanism. Check whether your deployment uses yamux as the stream muxer; if it does, consider whether you can switch to an alternative muxer or apply connection-level read timeouts as a temporary defense. Be aware that the attacker must complete the Noise handshake first, which may create a false sense of safety — authenticated peers should not be able to DoS each other at the protocol layer, and this vulnerability breaks that invariant. No patched version exists at the time of disclosure, which indicates the fix likely requires architectural changes rather than a simple patch.
CVE-2026-73568
Official description Straight from the sourceThe vendor's or NVD's own wording, published unedited. Authoritative, but often terse — it says what broke, rarely what to do.
NVD · uneditedpy-libp2p is the Python implementation of the libp2p networking stack. In 0.7.0 and earlier, the yamux handle_incoming() method in libp2p/stream_muxer/yamux/yamux.py reads an attacker-controlled 32-bit DATA frame length with read_exactly() before validating it against MAX_WINDOW_SIZE or checking whether stream_id exists. A peer that completes the standard Noise handshake can send a 12-byte frame declaring a 0xFFFFFFFF body and then withhold the body, causing the sequential yamux read loop used by the default new_host() configuration to block and preventing every stream on that connection from making progress. No fixed version is available as of this review.
Technical summary Written by usOur analysis, written from the advisory, the CVSS vector and the affected-version data. It adds context the advisory leaves out, and never invents facts that are not in the source.
dbcve analysisA single request can consume unbounded CPU, memory, or connections, so a modest amount of malicious traffic exhausts the service. The result is denial of service for everyone else. Remediation is enforcing limits, quotas, and timeouts on what any one request can use.
General guidance for the uncontrolled resource consumption class — the official description and references above are authoritative for this specific CVE. Want a bespoke review and a reviewed fix? Ask our team →
CVSS breakdown How the score is builtThe industry scoring standard. It rates how the flaw is reached, what it takes to exploit, and what an attacker gains — the score is derived from those, not the other way round.
From the vector- Attack vector
- Network
- Complexity
- Low
- Privileges
- None
- User interaction
- None
- Scope
- Unchanged
- Confidentiality
- None
- Integrity
- None
- Availability
- High
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Remediation Closing itWhat it takes to close this. Where a vendor fix exists we point at it; where none exists we say so plainly, and can build one. Effort estimates are scoped from the advisory, not from your codebase.
From vendor dataThere is no version to upgrade to and no patch to apply. Every affected install stays exposed until the vendor ships a fix — or somebody else builds one.
Free. We build fixes in the order the community asks for them — and we’ll tell you the moment this one lands.
We develop and verify an original fix where the vendor hasn’t, from $4,900. Deployed to your staging first — never straight to production.
Scope it with usSee what else the community needs solved on the solutions-needed board.
Scan for this in your stack
Free · runs locallyCheck whether your project pulls in CVE-2026-73568 — or any other known-vulnerable package — straight from your lock files. Free and open source; it runs locally and uploads nothing.
References Go to the primary sourcePrimary sources — vendor advisories, patches and trackers. Where our summary and a reference disagree, the reference wins.
Primary sourcespy-libp2p's yamux implementation contains a remote denial-of-service vulnerability that warrants your immediate attention. The issue is that py-libp2p reads the declared body length from a yamux frame header, then performs a blocking read for that exact number of bytes — and only AFTER the body arrives does it check whether the length exceeds MAX_WINDOW_SIZE. A malicious peer can send a frame with an arbitrarily large declared length, causing your process to block indefinitely waiting for data that will never arrive, or until TCP timeouts eventually free the connection. The attack works on the default multiplexer, meaning one 12-byte frame from an authenticated peer freezes every stream multiplexed over that connection simultaneously. The root cause is architectural: yamux was designed in Go with the assumption that blocking I/O is acceptable because goroutines are cheap. py-libp2p inherited this pattern without adapting it for Python's async context, creating a situation where a length field intended for flow control is being used as an implicit security boundary — which it isn't. The yamux spec treats MAX_WINDOW_SIZE as a buffer limit, not a DoS prevention mechanism. Check whether your deployment uses yamux as the stream muxer; if it does, consider whether you can switch to an alternative muxer or apply connection-level read timeouts as a temporary defense. Be aware that the attacker must complete the Noise handshake first, which may create a false sense of safety — authenticated peers should not be able to DoS each other at the protocol layer, and this vulnerability breaks that invariant. No patched version exists at the time of disclosure, which indicates the fix likely requires architectural changes rather than a simple patch.
Practitioner notes
ContributedPeer-ranked notes from engineers who’ve handled CVE-2026-73568 in production — separate from our analysis above.
The advisory tells you what broke. It rarely tells you what actually worked. If you’ve dealt with this one, that detail is what the next engineer is searching for.
- The version that genuinely resolved it — not the one the vendor claimed
- A config change or rule that shut the vector down
- A gotcha in the upgrade path that cost you an afternoon
A place for practitioners to share what actually worked: a mitigation you’ve tested, a configuration change, a version- or environment-specific caveat, or a link to a verified patch. The most useful notes rise to the top as peers upvote them, so the signal stays high.
- Verified mitigations, workarounds, and config changes
- Version or environment caveats, and links to real fixes
- No weaponised exploit code, or anything meant to cause harm
- No spam, self-promotion, credentials, or personal data