CVE-2026-41240

Summary of technical details (CVE-2026-41240) : DOMPurify prior to 3.4.0 has an inconsistency between FORBID_TAGS and FORBID_ATTR when function-based ADD_TAGS is used. The fix added an early exit for FORBID_ATTR but the FORBID_TAGS path remained unpatched, causing short-circuiting that allows for...

CVE-2026-23413

In the Linux kernel, the following vulnerability has been resolved: clsact: Fix use-after-free in init/destroy rollback asymmetry Fix a use-after-free in the clsact qdisc upon init/destroy rollback asymmetry. The latter is achieved by first fully initializing a clsact instance, and then in a seco...

CVE-2026-23413 clsact: Fix use-after-free in init/destroy rollback asymmetry

In the Linux kernel, the following vulnerability has been resolved: clsact: Fix use-after-free in init/destroy rollback asymmetry Fix a use-after-free in the clsact qdisc upon init/destroy rollback asymmetry. The latter is achieved by first fully initializing a clsact instance, and then in a seco...

PT-2026-29720

Name of the Vulnerable Software and Affected Versions Linux Kernel affected versions not specified Description A use-after-free issue exists in the clsact qdisc during init/destroy rollback asymmetry. This occurs when a clsact instance is fully initialized, and a subsequent replacement fails. The...

Handlebars.js has a Prototype Method Access Control Gap via Missing __lookupSetter__ Blocklist Entry

Summary The prototype method blocklist in lib/handlebars/internal/proto-access.js blocks constructor, defineGetter, defineSetter, and lookupGetter, but omits the symmetric lookupSetter. This omission is only exploitable when the non-default runtime option allowProtoMethodsByDefault: true is...

Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments

Large language models LLMs increasingly operate in multi-agent and safety-critical settings, raising open questions about how their vulnerabilities scale when models interact adversarially. This study examines whether larger models can systematically jailbreak smaller ones - eliciting harmful or...

Multimodal Safety Is Asymmetric: Cross-Modal Exploits Unlock Black-Box MLLMs Jailbreaks

Multimodal large language models MLLMs have demonstrated significant utility across diverse real-world applications. But MLLMs remain vulnerable to jailbreaks, where adversarial inputs can collapse their safety constraints and trigger unethical responses. In this work, we investigate jailbreaks i...

SUSE CVE-2022-50002

In the Linux kernel, the following vulnerability has been resolved: net/mlx5: LAG, fix logic over MLX5LAGFLAGNDEVSREADY Only set MLX5LAGFLAGNDEVSREADY if both netdevices are registered. Doing so guarantees that both ldev-pfMLX5LAGP0.dev and ldev-pfMLX5LAGP1.dev have valid pointers when...

Rewriting the Budget: a General Framework for Black-Box Attacks under Cost Asymmetry

Traditional decision-based black-box adversarial attacks on image classifiers aim to generate adversarial examples by slightly modifying input images while keeping the number of queries low, where each query involves sending an input to the model and observing its output. Most existing methods...

Rational Astrologies and Security

John Kelsey and I wrote a short paper for the Rossfest Festschrift: "Rational Astrologies and Security": There is another non-security way that designers can spend their security budget: on making their own lives easier. Many of these fall into the category of what has been called rational...

HTTP/2 Stream Cancellation Attack

HTTP/2 Rapid reset attack The HTTP/2 protocol allows clients to indicate to the server that a previous stream should be canceled by sending a RSTSTREAM frame. The protocol does not require the client and server to coordinate the cancellation in any way, the client may do it unilaterally. The clie...

Mitigation of M-12: Issue NOT mitigated

Mitigated issue M-12: No slippage protection on stake in SafEth.sol There were issues with either a lack of slippage protection or a hard set slippage. Slippage protection was missing in deposit for Reth.deposit only if depositing in the Rocket Pool and in Reth.withdraw, as well as in stake becau...

Mitigation Confirmed for M-02

MITIGATION IS NOT CONFIRMED MITIGATION IS NOT CONFIRMED Mitigation of M-02: Issue not mitigated Link to Issue: code-423n4/2023-03-asymmetry-findings1049 Comment Issue M-02 describes an edge case in which the SfrxEth derivative may revert under an scenario where the calculation of the redeem amoun...

Users might lose their stETH rebased reward due to the weights change

Lines of code Vulnerability details Vulnerability Details Let's consider the following scenario: Bob deposits 10eth with the weights: stETH: weights0 = 90e18, rETH: weights1 = 5e18. sfrxETH: weights2 = 5e18 Now, since the Lido has 80% of liquid staking market, Asymmetry Finance decides to adjust...

[H-01] RETH oracle manipulation allows attacker to steal funds

Lines of code Vulnerability details Impact The Asymmetry SafEth protocol aims to help diversify and decentralize liquid staking derivatives, exchanging ether staked in the protocol for staked ether derivative tokens based on some relative weighting. A function exposed by these derivative wrapper...

Buyer can take from a seller an item they had not offered

Lines of code Vulnerability details Impact Orders in Seaport can involve "criteria-based items", in which case it isn't one specific item that can fulfill the order but any one of a set of items. This set of items is the "criteria", and it is specified in the order as the merkle root of a merkle...

SSL / TLS Renegotiation DoS

The remote service encrypts traffic using TLS / SSL and permits clients to renegotiate connections. The computational requirements for renegotiating a connection are asymmetrical between the client and the server, with the server performing several times more work. Since the remote host does not...