llama.cpp before b11393 contains a use-after-free and double free vulnerability in...
Summary
Vulnerability reported (CVE-2026-107183). Successful exploitation can lead to full system takeover. Exploitable via POST /completion.
Original description
The description as published by the source.
llama.cpp before b11393 contains a use-after-free and double free vulnerability in common_chat_peg_mapper::map that allows unauthenticated remote attackers to corrupt heap memory via a dangling current_tool pointer. Attackers can submit a chat_parser in a POST /completion request emitting a tool-id after a tool-close tag to crash llama-server and shape a heap write primitive.
Source: NIST National Vulnerability Database NVD GitHub Advisory
Auto-generated summary
Assembled by SecPulse from the source data (CVSS, CWE, affected versions) using templates — no AI involved. Check the original description and references below for exact details.
A vulnerability tracked as CVE-2026-107183 has been found.
Attackers can target a specific entry point like POST /completion over the network to misuse the product.
Successful exploitation can lead to full system takeover. CVSS score: 8.1/10.
What to do: apply the vendor's official patched release.
If unsure, ask your IT team or search "CVE-2026-107183" on the vendor's site.
CVE-2026-107183 — CWE-416 / CVSS v3 8.1
Attack vector: remote (network-reachable) / unauthenticated / no user interaction
Attack surface: POST /completion
Plan: 1) Audit SBOM/dependencies, 2) Stage→prod upgrade, 3) Add WAF/proxy monitoring on affected endpoints, 4) Hunt IOCs in logs.
Refs: see the GHSA / vendor advisory / patched release linked on this page.
Use-after-free (CWE-416) has been reported. Attackers reach the vulnerable code path via POST /completion without authentication.
Check the vendor advisory for the affected versions. Attack surface: POST /completion.
Severity: High (CVSS 8.1/10). Successful exploitation can lead to full system takeover.
Update to the patched release per vendor advisory.
Until the patch is applied: disable the affected feature, apply WAF rules, or restrict access via network ACLs.
Search webserver/proxy logs for unusual POST /completion requests with malformed payloads or SQL meta-characters. Audit SBOM/dependencies to find affected services.
Response actions
Concrete steps and command examples for SOC/SRE teams to execute in order
-
1Identify exposure identify
Audit SBOM/dependencies for affected components.資産台帳・SBOM・依存マニフェストで影響コンポーネントを特定する。
-
2Hunt for indicators of compromise detect
grep 'POST /completion' /var/log/nginx/access.log | grep -E '(unusual_payload|sqli_pattern)'アクセスログで
POST /completionへの異常なリクエスト (不正な認証ヘッダ・SQLメタ文字)を過去 30〜90日分捜索。WAF/SIEM があれば該当パスのアラート発火履歴を確認。 -
3Post-deployment verification verify
Replay attack against POST /completion on staging to confirm patch closes the vectorパッチ適用後、ステージングで PoC または同等の悪用パターンを再現して脆弱性が閉じたことを確認。本番では Step 2 と同じログクエリでアラート再発が無いか継続監視。