GHSA-V76P-62QX-WWQ2
Vulnerability from github – Published: 2026-10-07 16:18 – Updated: 2026-10-07 16:18Summary
When decoding a tiled TIFF with fax compression (T4/T6/MH), DecodeTilesChunky allocates each tile buffer from TileWidth (ceil(TileWidth*bpp/8)*TileLength bytes) but constructs the fax decompressor with frame.Width: TiffDecompressorsFactory ignores the isTiled/tileWidth/tileHeight parameters entirely. The T4/T6/MH decompressors treat the full image width as the scanline length and advance (and really write, via read-modify-write bit ops) frame.Width bits per row, with no bounds check against the tile buffer. The very first tile therefore writes linearly out of bounds — about ImageWidth/8 bytes per row × TileLength rows into a TileWidth-sized buffer. With ImageWidth=4,000,000, TileWidth=16, TileLength=16 this writes ~2 MB past a 32-byte buffer and kills the process deterministically; a T6 all-white variant advances the bit offset by >512 MB silently, showing an alarm-free heap-corruption window for the same defect. A crafted file fully controls the OOB length per tile and works with perfectly legal per-row run codes (no overlong runs needed).
Verified at commit 5cd4d0d26a82a9549f297a237aea9cf665bddff8 (main; latest release v4.1.0, the supported major).
Details
Root cause: a size mismatch between tile buffer allocation and decompressor width, because the factory drops the tile parameters.
- Allocation: TiffDecoderCore.cs#L792-L794 —
bytesPerTileRow = RoundUpToMultipleOfEight(tileWidth*bitsPerPixel); tile buffer =bytesPerTileRow*tileLengthbytes (32 bytes in the PoC) - Mismatch: TiffDecoderCore.cs#L797 —
CreateDecompressor<TPixel>(frame.Width, ..., isTiled: true, tileWidth, tileLength)passes the full frame width - Factory drops tile params: TiffDecompressorsFactory.cs#L56-L64 — T4/T6/MH decompressors receive only
width(= frame.Width);isTiled/tileWidth/tileHeightignored - OOB write sink: T4TiffCompression.cs#L69-L119,
T6TiffCompression.cs#L76-L108— per-row advance ofthis.widthbits viaBitWriterUtils.WriteBitswith no buffer-length check (BitWriterUtils.cs#L51); noteTiffDecoderCore.cs:829-831later reads the buffer inbytesPerTileRowstrides, confirming the protocol expects tile-width rows
Attack surface: Image.Load(stream) on an attacker-supplied tiled TIFF (TiffDecoderCore.DecodeImageWithTiles → DecodeTilesChunky). Default configuration; only requirement is a standard tiled TIFF header (TileWidth=16, TileLength=16) + Compression=3 (T4; T6 also constructible).
Suggested remediation:
1. Short-term: when isTiled, construct the T4/T6/MH decompressor with tileWidth (not frame width), or clip row writes to the caller-provided buffer length.
2. Root fix: bound the write side of BitWriterUtils (pass remaining bits), and validate every fax row advance against buffer capacity on both tiled and strip paths.
3. Regression tests: Compression=2/3/4 × tiled with TileWidth < ImageWidth, including a T6 black-pixel row (forces real WriteBit).
PoC
Full PoC posted as the first comment below: Program.cs (driver), poc-tiled-t4.tif (crafted file, ~9.8 KB, base64 inline), README.
- Build a small console project referencing
src/ImageSharp/ImageSharp.csprojand run it against the crafted file (or callImage.Loadon it from any host). - Observed with ImageWidth=4,000,000, TileWidth=16, TileLength=16, T4 with 400 makeup codes + EOL per row:
tile payload 9642 bytes; rows write ~2,048,000 bytes into a 32-byte buffer
Fatal error. System.AccessViolationException: Attempted to read or write protected memory.
at SixLabors.ImageSharp.Formats.Tiff.Compression.BitWriterUtils.WriteBits(Span`1<Byte>, IntPtr, IntPtr, Byte)
at ...T4TiffCompression.WritePixelRun(...)
at ...T4TiffCompression.Decompress(...)
Aborted (core dumped); exit=134
- Controls: the same file with Compression=None decodes normally (container is fine); a T6 all-white-rows variant advances >512 MB of bit offset without a real write (silent corruption window) before tripping on a directory-level TileOffsets count check.
Impact
- What it is: out-of-bounds write (CWE-787). For any service decoding untrusted tiled TIFFs: remote, default-configuration, deterministic process crash (DoS), plus a heap OOB write whose per-tile length and row width are attacker-tunable — a potential code-execution surface. This is a vulnerability in the library itself, in scope of your SECURITY.md.
- Who is impacted: applications decoding untrusted TIFF with SixLabors.ImageSharp at the current major (verified on main past v4.1.0); tiled + fax-compressed files are the trigger, which ordinary TIFF writers can produce.
Reported by Kimi Security Team (bug-report@moonshot.ai).
{
"affected": [
{
"package": {
"ecosystem": "NuGet",
"name": "ImageSharp"
},
"ranges": [
{
"events": [
{
"introduced": "3.0.0"
},
{
"fixed": "4.1.1"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-106118"
],
"database_specific": {
"cwe_ids": [
"CWE-787"
],
"github_reviewed": true,
"github_reviewed_at": "2026-10-07T16:18:57Z",
"nvd_published_at": "2026-10-06T19:17:42Z",
"severity": "HIGH"
},
"details": "## Summary\n\nWhen decoding a tiled TIFF with fax compression (T4/T6/MH), `DecodeTilesChunky` allocates each tile buffer from **TileWidth** (`ceil(TileWidth*bpp/8)*TileLength` bytes) but constructs the fax decompressor with **frame.Width**: `TiffDecompressorsFactory` ignores the `isTiled/tileWidth/tileHeight` parameters entirely. The T4/T6/MH decompressors treat the full image width as the scanline length and advance (and really write, via read-modify-write bit ops) `frame.Width` bits per row, with no bounds check against the tile buffer. The very first tile therefore writes linearly out of bounds \u2014 about `ImageWidth/8` bytes per row \u00d7 TileLength rows into a `TileWidth`-sized buffer. With ImageWidth=4,000,000, TileWidth=16, TileLength=16 this writes ~2 MB past a 32-byte buffer and kills the process deterministically; a T6 all-white variant advances the bit offset by \u003e512 MB silently, showing an alarm-free heap-corruption window for the same defect. A crafted file fully controls the OOB length per tile and works with perfectly legal per-row run codes (no overlong runs needed).\n\nVerified at commit `5cd4d0d26a82a9549f297a237aea9cf665bddff8` (main; latest release v4.1.0, the supported major).\n\n## Details\n\nRoot cause: a size mismatch between tile buffer allocation and decompressor width, because the factory drops the tile parameters.\n\n- Allocation: [TiffDecoderCore.cs#L792-L794](https://github.com/SixLabors/ImageSharp/blob/5cd4d0d26a82a9549f297a237aea9cf665bddff8/src/ImageSharp/Formats/Tiff/TiffDecoderCore.cs#L792-L794) \u2014 `bytesPerTileRow = RoundUpToMultipleOfEight(tileWidth*bitsPerPixel)`; tile buffer = `bytesPerTileRow*tileLength` bytes (32 bytes in the PoC)\n- Mismatch: [TiffDecoderCore.cs#L797](https://github.com/SixLabors/ImageSharp/blob/5cd4d0d26a82a9549f297a237aea9cf665bddff8/src/ImageSharp/Formats/Tiff/TiffDecoderCore.cs#L797) \u2014 `CreateDecompressor\u003cTPixel\u003e(frame.Width, ..., isTiled: true, tileWidth, tileLength)` passes the full frame width\n- Factory drops tile params: [TiffDecompressorsFactory.cs#L56-L64](https://github.com/SixLabors/ImageSharp/blob/5cd4d0d26a82a9549f297a237aea9cf665bddff8/src/ImageSharp/Formats/Tiff/Compression/TiffDecompressorsFactory.cs#L56-L64) \u2014 T4/T6/MH decompressors receive only `width` (= frame.Width); `isTiled/tileWidth/tileHeight` ignored\n- OOB write sink: [T4TiffCompression.cs#L69-L119](https://github.com/SixLabors/ImageSharp/blob/5cd4d0d26a82a9549f297a237aea9cf665bddff8/src/ImageSharp/Formats/Tiff/Compression/Decompressors/T4TiffCompression.cs#L69-L119), `T6TiffCompression.cs#L76-L108` \u2014 per-row advance of `this.width` bits via `BitWriterUtils.WriteBits` with no buffer-length check ([BitWriterUtils.cs#L51](https://github.com/SixLabors/ImageSharp/blob/5cd4d0d26a82a9549f297a237aea9cf665bddff8/src/ImageSharp/Formats/Tiff/Compression/BitWriterUtils.cs#L51)); note `TiffDecoderCore.cs:829-831` later reads the buffer in `bytesPerTileRow` strides, confirming the protocol expects tile-width rows\n\nAttack surface: `Image.Load(stream)` on an attacker-supplied tiled TIFF (`TiffDecoderCore.DecodeImageWithTiles` \u2192 `DecodeTilesChunky`). Default configuration; only requirement is a standard tiled TIFF header (TileWidth=16, TileLength=16) + Compression=3 (T4; T6 also constructible).\n\n**Suggested remediation:**\n1. Short-term: when `isTiled`, construct the T4/T6/MH decompressor with `tileWidth` (not frame width), or clip row writes to the caller-provided buffer length.\n2. Root fix: bound the write side of `BitWriterUtils` (pass remaining bits), and validate every fax row advance against buffer capacity on both tiled and strip paths.\n3. Regression tests: Compression=2/3/4 \u00d7 tiled with TileWidth \u003c ImageWidth, including a T6 black-pixel row (forces real `WriteBit`).\n\n## PoC\n\nFull PoC posted as the first comment below: `Program.cs` (driver), `poc-tiled-t4.tif` (crafted file, ~9.8 KB, base64 inline), README.\n\n1. Build a small console project referencing `src/ImageSharp/ImageSharp.csproj` and run it against the crafted file (or call `Image.Load` on it from any host).\n2. Observed with ImageWidth=4,000,000, TileWidth=16, TileLength=16, T4 with 400 makeup codes + EOL per row:\n\n ```\n tile payload 9642 bytes; rows write ~2,048,000 bytes into a 32-byte buffer\n Fatal error. System.AccessViolationException: Attempted to read or write protected memory.\n at SixLabors.ImageSharp.Formats.Tiff.Compression.BitWriterUtils.WriteBits(Span`1\u003cByte\u003e, IntPtr, IntPtr, Byte)\n at ...T4TiffCompression.WritePixelRun(...)\n at ...T4TiffCompression.Decompress(...)\n Aborted (core dumped); exit=134\n ```\n\n3. Controls: the same file with Compression=None decodes normally (container is fine); a T6 all-white-rows variant advances \u003e512 MB of bit offset without a real write (silent corruption window) before tripping on a directory-level TileOffsets count check.\n\n## Impact\n\n- **What it is:** out-of-bounds write (CWE-787). For any service decoding untrusted tiled TIFFs: remote, default-configuration, deterministic process crash (DoS), plus a heap OOB write whose per-tile length and row width are attacker-tunable \u2014 a potential code-execution surface. This is a vulnerability in the library itself, in scope of your SECURITY.md.\n- **Who is impacted:** applications decoding untrusted TIFF with SixLabors.ImageSharp at the current major (verified on main past v4.1.0); tiled + fax-compressed files are the trigger, which ordinary TIFF writers can produce.\n\n---\n\n---\n\nReported by **Kimi Security Team** (bug-report@moonshot.ai).",
"id": "GHSA-v76p-62qx-wwq2",
"modified": "2026-10-07T16:18:57Z",
"published": "2026-10-07T16:18:57Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/SixLabors/ImageSharp/security/advisories/GHSA-v76p-62qx-wwq2"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-106118"
},
{
"type": "WEB",
"url": "https://github.com/SixLabors/ImageSharp/pull/3176"
},
{
"type": "WEB",
"url": "https://github.com/SixLabors/ImageSharp/commit/9ee7d1dd5b62d8c9f16cc755e76828386bc2191f"
},
{
"type": "PACKAGE",
"url": "https://github.com/SixLabors/ImageSharp"
},
{
"type": "WEB",
"url": "https://github.com/SixLabors/ImageSharp/releases/tag/v4.1.1"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "ImageSharp: Tiled fax TIFF: tile buffer sized by TileWidth but fax decompressor writes scanlines of ImageWidth \u2014 heap OOB write"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.