GitHub Advisory Database
Security vulnerability database inclusive of CVEs and GitHub originated security advisories from the world of open source software.
Filter advisories
GitHub reviewed advisories
Unreviewed advisories
Malware advisories
Filter advisories
Filter advisories
GitHub reviewed advisories
All reviewed
5,000+
Composer
5,000+
Erlang
108
GitHub Actions
54
Go
4,507
Maven
5,000+
npm
5,000+
NuGet
1,091
pip
5,000+
Pub
13
RubyGems
1,144
Rust
1,511
Swift
62
Unreviewed advisories
All unreviewed
5,000+
Malware advisories
All malware
5,000+
Composer
1
Go
18
Maven
2
npm
5,000+
NuGet
264
pip
5,000+
RubyGems
3,510
Rust
10
42 advisories
Filter by severity
vLLM: Processing differential in multi-channel audio downmixing enables hidden-input/moderation bypass for audio models
Moderate
CVE-2026-34760
was published
for
vllm
(pip)
Jul 17, 2026
vLLM: OOM Denial of Service via Audio Decompression Bomb
Moderate
CVE-2026-54233
was published
for
vllm
(pip)
Jun 17, 2026
vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant serving
Moderate
CVE-2026-53923
was published
for
vllm
(pip)
Jun 17, 2026
vLLM: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations
Moderate
CVE-2026-12491
was published
for
vllm
(pip)
Jun 17, 2026
vLLM's Artifact Pin Decay allows pinned deployments to load unpinned code, weights, and processors
Moderate
CVE-2026-47155
was published
for
vllm
(pip)
Jun 10, 2026
vLLM: Denial of Service via Unbounded Frame Count in video/jpeg Base64 Processing
Moderate
CVE-2026-34755
was published
for
vllm
(pip)
Apr 3, 2026
vLLM: Server-Side Request Forgery (SSRF) in `download_bytes_from_url `
Moderate
CVE-2026-34753
was published
for
vllm
(pip)
Apr 3, 2026
vLLM: Unauthenticated OOM Denial of Service via Unbounded `n` Parameter in OpenAI API Server
Moderate
CVE-2026-34756
was published
for
vllm
(pip)
Apr 3, 2026
vLLM has Hardcoded Trust Override in Model Files Enables RCE Despite Explicit User Opt-Out
High
CVE-2026-27893
was published
for
vllm
(pip)
Mar 27, 2026
vLLM vulnerable to Server-Side Request Forgery (SSRF) through MediaConnector
High
CVE-2026-24779
was published
for
vllm
(pip)
Jan 28, 2026
vLLM affected by RCE via auto_map dynamic module loading during model initialization
High
CVE-2026-22807
was published
for
vllm
(pip)
Jan 21, 2026
vLLM vulnerable to remote code execution via transformers_utils/get_config
High
CVE-2025-66448
was published
for
vllm
(pip)
Dec 2, 2025
vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs`
Moderate
CVE-2025-62426
was published
for
vllm
(pip)
Nov 20, 2025
vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs
High
CVE-2025-62372
was published
for
vllm
(pip)
Nov 20, 2025
vLLM deserialization vulnerability leading to DoS and potential RCE
High
CVE-2025-62164
was published
for
vllm
(pip)
Nov 20, 2025
vLLM is vulnerable to Server-Side Request Forgery (SSRF) through `MediaConnector` class
High
CVE-2025-6242
was published
for
vllm
(pip)
Oct 7, 2025
vLLM is vulnerable to timing attack at bearer auth
High
CVE-2025-59425
was published
for
vllm
(pip)
Oct 7, 2025
vLLM has remote code execution vulnerability in the tool call parser for Qwen3-Coder
High
CVE-2025-9141
was published
for
vllm
(pip)
Aug 21, 2025
vllm API endpoints vulnerable to Denial of Service Attacks
High
CVE-2025-48956
was published
for
vllm
(pip)
Aug 21, 2025
vLLM Tool Schema allows DoS via Malformed pattern and type Fields
Moderate
CVE-2025-48944
was published
for
vllm
(pip)
May 28, 2025
vLLM allows clients to crash the openai server with invalid regex
Moderate
CVE-2025-48943
was published
for
vllm
(pip)
May 28, 2025
vLLM DOS: Remotely kill vllm over http with invalid JSON schema
Moderate
CVE-2025-48942
was published
for
vllm
(pip)
May 28, 2025
ProTip!
Advisories are also available from the
GraphQL API