ImageSharp CVE-2026-106111

Exploitation requires the target application to decode an attacker-supplied EXR image that uses ZIP or ZIPS compression (ExrBaseDecompressor.UndoZipCompression) with ImageSharp versions 4.0.0 through 4.1.1, AND to then expose the decoded pixels or output back to the attacker. The attacker must craft the EXR so the inflate result is shorter than the EXR block's required size. The disclosure is only meaningful when the reused Configuration.Default pooled buffer still contains bytes from a completed prior ImageSharp operation - the attacker cannot directly control that residual content, which is the main limiting factor (reflected as AC:H). Exploitation does not require authentication (PR:N) or user interaction (UI:N), and EXR decoding is a supported default code path, but leaked data is non-deterministic and limited to confidentiality (no integrity or availability impact).

This is a genuine but moderate-priority information-disclosure flaw, consistent with its 5.9 (Medium) CVSS score rather than an urgent patch-now emergency. The CVSS vector CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:N/A:N shows a network-reachable, unauthenticated issue with HIGH attack complexity and confidentiality-only impact (CWE-226, sensitive information not removed before buffer reuse). The high complexity is well-justified by the mechanism: exploitation depends on residual bytes from a completed prior ImageSharp operation still being present in a buffer rented from Configuration.Default, which the attacker cannot deterministically control - it requires favorable heap/pool state and timing. Exploitation status: no public exploit identified at time of analysis. No EPSS score, KEV listing, or POC was provided in the input, which is consistent with a low likelihood of widespread, automated exploitation; this looks like a niche, opportunistic data-leak rather than a mass-exploitable RCE. No conflicting signals were observed, but note that the provided source-code diff (ICC channel-count bounds checks and stream extension refactoring) does NOT visibly contain the described EXR/ZIP decompressor fix, so the exact code change could not be independently confirmed from the supplied diff.

An attacker submits a specially crafted EXR image whose ZIP/ZIPS-compressed block inflates to fewer bytes than the EXR block's declared size to an application that decodes untrusted images with ImageSharp and then returns or displays the decoded pixels. Because the short inflate result is reconstructed over a pooled buffer that still holds data from a previous ImageSharp operation, stale process-local image bytes bleed into the decoded output and are disclosed to the attacker. No public exploit code is identified, and the high CVSS attack complexity reflects the need for exploitable residual data to be present in the reused buffer.