GitLab Advisory Database (GLAD)

When the vLLM API receives a malformed request (e.g., invalid JSON or missing required fields), FastAPI raises a Pydantic RequestValidationError. The validation_exception_handler in vllm/entrypoints/openai/server_utils.py converts this exception to a string via str(exc), which includes the internal file path and line number of the handler function. The existing sanitize_message() function in vllm/entrypoints/utils.py strips memory addresses (e.g., 0x7f…) but does not strip File "…", line X patterns. The result is a user-facing …

The fix for GHSA-rwxx-mrjm-wc2m ("ReDoS via structured_outputs.regex compiled without timeout") wrapped the regex compile in the xgrammar and outlines backends with compile_regex_with_timeout (and, for outlines, validate_regex_is_buildable). The lm-format-enforcer backend was left unguarded: it compiles the attacker-supplied regex with no timeout and no buildability check. A single request with a catastrophic regex hangs the structured-output compile step and stalls the engine worker (denial of service).

The /v1/completions/derender and /v1/chat/completions/derender endpoints accept caller-supplied GenerateResponse objects and postprocess every nested choices[*].token_ids list directly. Unlike the normal render/generate path, derender does not enforce model context length, resolved max_tokens, max_num_seqs, choice-count, or response-size bounds before detokenizing and returning the supplied token IDs. An authenticated API client can therefore make the CPU-only render frontend, or any server exposing these /v1 derender routes, spend CPU and memory proportional to attacker-chosen generated-output-shaped …

@typespec/spector registers a POST /.admin/stop HTTP route with no authentication, authorization token, Origin check, or IP-source restriction. Any network-reachable client can send a single unauthenticated POST request to terminate the mock server process. Because the server binds to 0.0.0.0 by default (all interfaces), this endpoint is exposed to any host that can reach the server's port—not just localhost—making a complete denial-of-service trivially achievable with one HTTP request. Severity is High …

A PERMISSIONS … WHERE clause is evaluated with permission enforcement disabled, so it can't recurse into its own checks. But the clause could also contain data-modifying statements, and these ran with enforcement still off — so evaluating a permission check could write to tables the caller cannot write. For example: DEFINE TABLE post PERMISSIONS FOR update WHERE (CREATE log SET at = time::now()) OR true; Any user allowed to update …

An authenticated user scoped to one namespace/database could invoke a custom API (DEFINE API) belonging to a different namespace/database, reaching another tenant's endpoint. The route /api/{namespace}/{database}/{endpoint} took the namespace and database from the URL and applied them to the caller's session before the endpoint was looked up or run, without checking that the caller's authenticated scope covered them. Because a custom API handler runs with permissions disabled (definer's rights), the …

SurrealDB offers http functions that can access external network endpoints. A typical, albeit not recommended configuration would be to start SurrealDB with all network connections allowed with the exception of a deny list. For example, surreal start –allow-net –deny-net 10.0.0.0/8 will allow all network connections except to the 10.0.0.0/8 block. An authenticated user of SurrealDB can use bypass this restriction, using http::<fn>(<url>) functions where the hostname resolves to an IP …

/api/system/getConf returns Conf.CookieKey, the key used to sign the server's session cookies in its response body. The endpoint is registered with CheckAuth only, so the field reaches the publish RoleReader token and the anonymous account when Publish.Auth.Enable is false. The configuration-export endpoint in the same file strips this exact field before returning config, so the project already treats it as secret. The reader-facing masking path does not.

Insufficient clearing of the output buffer in Java-based decompressor implementations in lz4-java 1.10.0 and earlier allows remote attackers to read previous buffer contents via crafted compressed input. In applications where the output buffer is reused without being cleared, this may lead to disclosure of sensitive data. JNI-based implementations are not affected.

Insufficient clearing of the output buffer in Java-based decompressor implementations in lz4-java 1.10.0 and earlier allows remote attackers to read previous buffer contents via crafted compressed input. In applications where the output buffer is reused without being cleared, this may lead to disclosure of sensitive data. JNI-based implementations are not affected.

Insufficient clearing of the output buffer in Java-based decompressor implementations in lz4-java 1.10.0 and earlier allows remote attackers to read previous buffer contents via crafted compressed input. In applications where the output buffer is reused without being cleared, this may lead to disclosure of sensitive data. JNI-based implementations are not affected.

This advisory has been marked as a false positive.

This advisory has been marked as a false positive.

This advisory has been marked as a false positive.

This advisory has been marked as False Positive and has been removed.

After an API Token exposure from an exploited trivy dependency, two new releases of litellm were uploaded to PyPI containing automatically activated malware, harvesting sensitive credentials and files, and exfiltrating to a remote API. Anyone who has installed and run the project should assume any credentials available to litellm environment may have been exposed, and revoke/rotate thema ccordingly.