CVE WatchJun 30, 2026

CVE-2026-55698

pnpm is a package manager. Prior to 10.34.2 and 11.5.3, pnpm can persist package-manager bootstrap metadata in the first YAML document of pn

CVSS

8.8

High

EPSS

0.3%

p22

KEV

—

Exploit Today

0-100

Published: Jun 25, 2026 · Last modified: Jun 30, 2026 · CWE-345 · CWE-494 · CWE-829

EPSS · 30d

0.3%EPSS · 30 days0.3%

2026-08-112026-09-07

Technical description

pnpm is a package manager. Prior to 10.34.2 and 11.5.3, pnpm can persist package-manager bootstrap metadata in the first YAML document of pnpm-lock.yaml. Before the patch, direct pnpm execution trusted an already resolved packageManagerDependencies entry when the committed env lockfile contained matching pnpm and @pnpm/exe versions. A malicious repository could therefore commit package-manager lockfile package records and snapshots that bypassed fresh package-manager resolution, then cause pnpm to install and execute bytes selected by that committed lockfile state during automatic version switching. This vulnerability is fixed in 10.34.2 and 11.5.3.

Official references

Related CVEs

CVECVSSEPSSKEVRExploitTitleMod.

CVE-2026-79721—

—

———Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.9h CVE-2026-733167.5 HIG

—

———XenForo before 2.3.13 contains a payment replay vulnerability in the PayPal REST payment provider that allows attackers to process the same webhook payload multiple times by exploiting a missing duplicate transaction ID check. Attackers can replay a valid webhook payload to trigger duplicate payment events, resulting in repeated subscription activations and unauthorized account upgrades.9h CVE-2026-626546.8 MED

—

——0A vulnerability has been identified in Reyrolle 7SR5 (All versions < V2.70). A special maintenance mode can be activated via a physical key sequence during device boot, in which the device downloads and executes program code from a network server without verifying its authenticity or integrity. This could allow an attacker with physical access to the device to upload and execute arbitrary, unsigned code.10h CVE-2026-865047.8 HIG

—

——0In JetBrains IntelliJ IDEA before 2026.2.2 missing project-trust confirmation before building a Dev Container allowed host-level code execution13h CVE-2026-861698.8 HIG

39.9%

——12Axolotl through 0.18.0 contains a remote code execution vulnerability in the multipack patch path where trust_remote_code defaults to None instead of False, causing the security guard to be bypassed. Attackers can execute arbitrary Python code by crafting a malicious Hugging Face model repository selected as base_model, which is loaded with hardcoded trust_remote_code=True during AutoModelForCausalLM.from_pretrained.9h CVE-2026-850083.7 LOW

1.8%

——1undici's cache interceptor documents that only safe HTTP methods are cached, but its logic to skip caching is built by subtracting the configured methods from the set of safe methods, so an unsafe method such as POST, PUT, or DELETE is never placed in the skip list and instead falls through to the full cache-read path. The response-storage gate also lacked a method check, so a response to an unsafe request that is heuristically cacheable or carries an explicit Cache-Control directive is stored and later replayed from cache. Because response headers from a remote origin are untrusted, an origin can answer once with a cacheable status and then have the client's own subsequent state-changing requests to that path served from the stale cache entry without ever reaching the origin, an integrity failure that occurs under the interceptor's default configuration. This affects undici versions from 7.0.0 up to 7.29.1 and from 8.0.0 up to 8.10.2. Users should upgrade to undici 7.29.1 or 8.10.2.9h