SGLang through 0.5.21 contains a resource budget bypass vulnerability that allows remote attackers to override GLM-V video processing limits by supplying fps, maxframes, maxtokensperframe and maximagetokens in the videourl object. Attackers can set these values arbitrarily high so glm4v.py decodes every frame of high-resolution video at full size, exhausting worker memory and denying service.

SGLang through 0.5.21 contains an unbounded memory allocation vulnerability in preprocessvideo in qwenvl.py that decodes frames at full source resolution before applying maxpixels limits. Remote attackers can submit high-resolution videos to /v1/chat/completions on Qwen-VL deployments, and concurrent requests exhaust worker memory causing denial of service.

SGLang through 0.5.21 contains a denial of service vulnerability that disables Pillow decompression bomb protection process-wide by setting Image.MAXIMAGEPIXELS to None in NanoNemotronVLImageProcessor. Remote attackers can submit small, highly compressed images with huge dimensions to the generic image loader, exhausting worker memory during decoding.

SGLang through 0.5.21 contains an unbounded memory allocation vulnerability that allows unauthenticated remote attackers to exhaust memory by uploading oversized files to /v1/audio/transcriptions. Attackers can send concurrent large multipart uploads, which are fully read into memory before validation, to crash the HTTP/tokenizer worker and deny service.

SGLang through 0.5.21 contains a decompression amplification vulnerability in loadaudio that allows remote attackers to exhaust memory by submitting highly compressed audio decoded without duration or sample limits. Attackers can supply small FLAC-encoded silence files through audiourl or audiodata inputs that expand into massive PCM buffers, exhausting worker memory and CPU for denial of service.

SGLang contains an arbitrary code execution vulnerability caused by the ZMQ message decoder unconditionally deserializing PickleWrapper payloads via pickle.loads() in maybeunwrappickle without type allowlisting or authentication; this vulnerability persists via the msgpack path even when SGLANGUSEPICKLEIPC is disabled, and becomes remotely exploitable if data-parallel attention is enabled with a non-loopback --dist-init-addr setting.

SGLang through 0.5.20 in prefill/decode disaggregation mode fails to validate duplicate bootstraproom fields in /generate requests with Mooncake KV transfer backend. Unauthenticated attackers can send concurrent requests with identical bootstraproom values to crash scheduler processes or hang other users' requests until transfer timeout.

SGLang's multimodal generation runtime is vulnerable to unauthenticated arbitrary code execution because the disaggregated-diffusion orchestrator's DiffusionServer binds an unauthenticated ZeroMQ ROUTER socket to a network interface and passes the final frame of received multipart messages directly to pickle.loads() before any validation occurs.

SGLang contains a DoS vulnerability caused by missing input validation for AUXDATA ZeroMQ control messages in the Decode worker, which enables an unauthenticated remote attacker with network reachability to the Decode control PULL socket to terminate the Decode control thread and cause a denial of service against the target server.

SGLang contains an SSRF and local file read in the multimodal generation endpoint /v1/chat/completions due to unsanitized imageurl, allowing access to internal metadata, secrets, and services.