b10451

ggml-org/llama.cppb10451Aug 16, 2026by github-actions[bot]same day as b10448

What changed since b10448

low risk

Patch release adds file-bounds validation for LoRA tensor data; no other changes since b10448.

Improvements

  • LoRA adapter loading now validates that tensor data offsets and sizes stay within the adapter file, hardening against malformed inputs.

Bug fixes

  • Malformed or truncated LoRA files can no longer cause out-of-bounds reads or crashes during adapter application.

Upgrade notes

This is a drop-in patch with no config, CLI, or model-format changes; existing LoRA adapters that load correctly today will continue to work. Only adapters with corrupt or truncated tensor data will now be rejected instead of crashing. No migration steps are required.

Analyzed Oct 24, 2026

Upgrading?

Release notes summary

This release adds a safety check to ensure LoRA tensor data is within the bounds of the file to prevent potential crashes from malformed LoRA files.

Key Highlights

  • Added bounds checking for LoRA tensor data
  • Prevents crashes from malformed LoRA files

New Features

  • LoRA file bounds validation

Release Notes (excerpt)

<details open> llama : check LoRA tensor data is within file bounds (#27056) * llama : check LoRA tensor data is within file bounds * Update src/llama-adapter.cpp Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co> --------- Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co> </details> **Website:** - <https://llama.app> **macOS/iOS:** - macOS Apple Silicon (arm64) - macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED - macOS Intel (x64)…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories