71 matches found
vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`
Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the flash late-interaction query cache reaches. Summary On late-interaction /score and /rerank deployments...
GHSA-2PHQ-3PHC-84PX vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`
Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the flash late-interaction query cache reaches. Summary On late-interaction /score and /rerank deployments...
EUVD-2026-92845
vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on /score and /rerank...
CVE-2026-105755
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's querykey value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can...
CVE-2026-105755
The vLLM inference and serving engine (prior to version 0.30.0 ) is vulnerable to a cross-request integrity break in its flash late-interaction scoring at the /score and /rerank endpoints. The root cause is that the system derives the query_key value from the caller-controlled X-Request-Id header...
CVE-2026-105755 vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's querykey value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can...
CVE-2026-105755 vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id — cross-request integrity break and induced errors on `/score` and `/rerank`
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, flash late-interaction scoring at the /score and /rerank endpoints derives each worker's querykey value from the caller-controlled X-Request-Id header. A concurrent request that reuses a victim's identifier can...
PT-2026-106265
Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.30.0 Description In deployments using flash late-interaction scoring, the engine derives worker cache keys for query embeddings from the X-Request-Id header, which is controlled by the caller. A concurrent request that...
PT-2026-93260
A vulnerability was found in vllm-project vLLM 0.26.0/0.27.0. Affected is the function MoRIIOConnectorScheduler.request finished/MoRIIOConnectorWorker.get finished/MoRIIOWrapper. handle release message of the file vllm/distributed/kv transfer/kv connector/v1/moriio/moriio connector.py of the...
libssh: libssh: denial of service via SFTP responses with unknown request IDs
A flaw was found in libssh. A malicious SFTP server can send responses for unknown request IDs that libssh clients keep queued indefinitely, causing unbounded memory growth and client-side denial of service...
PT-2026-70843
CVAT is an open source interactive video and image annotation tool for computer vision. From 2.17.0 until 2.72.0, a user with the Worker role can use predictable task-based request IDs with the lambda request retrieve and destroy endpoints to view automatic annotation requests for tasks or jobs t...
CVE-2026-17556 Path traversal in GitHub Enterprise Server allowed unauthenticated deletion of instance storage via the X-GitHub-Request-Id header
A path traversal vulnerability was identified in GitHub Enterprise Server that allowed an unauthenticated attacker to delete arbitrary files and directories on the instance, including the entire user storage directory containing Git LFS objects, release assets, attachments, and avatars. The...
CVE-2026-17556 Path traversal in GitHub Enterprise Server allowed unauthenticated deletion of instance storage via the X-GitHub-Request-Id header
A path traversal vulnerability was identified in GitHub Enterprise Server that allowed an unauthenticated attacker to delete arbitrary files and directories on the instance, including the entire user storage directory containing Git LFS objects, release assets, attachments, and avatars. The...
CVE-2026-17556: Improper Limitation of a Pathname to a Restricted Directory
A path traversal vulnerability was identified in GitHub Enterprise Server that allowed an unauthenticated attacker to delete arbitrary files and directories on the instance, including the entire user storage directory containing Git LFS objects, release assets, attachments, and avatars. The...
OPENSUSE-SU-2026:21444-1 Security update for libssh
This update for libssh fixes the following issues: - CVE-2026-15370: stack buffer overflow in SFTP server longname construction bsc1272162. - CVE-2026-59843: denial of service via zero advertised channel packet size bsc1272164. - CVE-2026-59844: denial of service via oversized SFTP read length...
CVE-2026-54602
FastGPT is a knowledge-based AI application platform. Prior to 4.15.0, GET /api/core/ai/record/getRecord authenticates the caller but loads LLM request and response traces only by requestId without team scoping, allowing any authenticated user to read another team's prompts, retrieved RAG chunks,...
CVE-2026-54602 FastGPT: Cross-team LLM request/response disclosure (IDOR) via /api/core/ai/record/getRecord
FastGPT is a knowledge-based AI application platform. Prior to 4.15.0, GET /api/core/ai/record/getRecord authenticates the caller but loads LLM request and response traces only by requestId without team scoping, allowing any authenticated user to read another team's prompts, retrieved RAG chunks,...
PT-2026-56260
Name of the Vulnerable Software and Affected Versions FastGPT versions prior to 4.15.0 Description An issue exists where the endpoint "/api/core/ai/record/getRecord" authenticates the caller but fails to implement team scoping when loading LLM request and response traces. By providing a known...
Unity Linux 20.1060a / 20.1070a Security Update: kernel (UTSA-2025-993054)
The Unity Linux 20 host has a package installed that is affected by a vulnerability as referenced in the UTSA-2025-993054 advisory. In the Linux kernel, the following vulnerability has been resolved: scsi: zfcp: Fix double free of FSF request when qdio send fails We used to use the wrong type of...
CVE-2025-66553
Summary: Nextcloud Tables prior to 0.8.7 and 0.9.4 allows authenticated users to view column metadata of other tables by altering the numeric ID in a request, causing information disclosure. The issue is fixed in 0.8.7 and 0.9.4. Remediation: upgrade Nextcloud Tables to version 0.8.7 or later, or...