Advisory Catalog

vLLM: SSRF + arbitrary local file read in MiMoV2OmniMultiModalProcessor `_fetch_image` and audio loader bypass MediaConnector protections

vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds

sqlparse: Inefficient Regex Handling of Dollar-Quoted SQL Literals Leads to ReDoS (Denial of Service)

sqlparse: TokenList.__init__ materializes O(subtree) value per group, causing CPU DoS before depth/token caps trigger

sqlparse: Quadratic O(n²) DoS in group_comments

Glances: as_dict_secure() Value-Level Bypass Leaks Credentials in URL Values via /api/4/config

Glances: `--disable-config-exec` does not cover on-alert action commands (incomplete fix of CVE-2026-53925)

Glances: Incomplete fix of CVE-2026-32608: action-template sanitizer is bypassed by nested stat values (process 'cmdline') → OS command injection

sqlparse: Generated Python and PHP snippets allow SQL string breakout through unescaped backslashes

Glances: REST API CORS Credentials Guard Uses Exact-Match Instead of Membership Test — Bypassed by Any Multi-Origin Allowlist Containing the Wildcard

Glances has a command injection bypass of action-template sanitizer via cross-field shell-operator reconstruction

openssl_encrypt versions before 1.4.0 contain an authentication bypass vulnerability in pqc.py where AES-GCM decryption failures trigger fallback to unauthenticated AES-CTR mode. Attackers can modify ciphertext in transit to bypass integrity verification and perform bit-flipping attacks without dete

openssl_encrypt versions before 1.4.0 contain a critical vulnerability in pqc.py where KEM decapsulation failures silently fall back to simulation mode, generating a deterministic shared secret from only 16 bytes of the private key and publicly available encapsulated key data. Attackers who obtain 1

openssl_encrypt versions before 1.4.0 contain a sandbox escape vulnerability in IsolatedPluginExecutor that exposes Python type objects in restricted exec() builtins. Attackers can traverse the Python class hierarchy via __class__.__mro__.__subclasses__() to access system functions and execute arbit

openssl_encrypt versions before 1.4.0 contain a sandbox escape vulnerability in the DangerousPatternVisitor AST analyzer that fails to detect dunder attribute traversal techniques. Attackers can use __class__, __bases__, __subclasses__(), and __globals__ chains to access restricted functions and exe

openssl_encrypt versions before 1.4.0 fail to apply sandbox restrictions in the default process isolation mode for plugin execution. Attackers can execute malicious plugins with unrestricted access to the filesystem, network, subprocess execution, and all Python modules.

openssl_encrypt before 1.4.0 contains an authentication bypass vulnerability in the verify_api_token function that accepts any non-empty Bearer token string without validation. Attackers can upload arbitrary public keys, enumerate all keys, and revoke keys belonging to any user by providing any Bear

openssl_encrypt versions before 1.4.0 contain hardcoded default JWT signing secrets in config.py that pass validation checks. Attackers with access to source code can forge valid JWT tokens for any client_id to gain authenticated access to keyserver and telemetry APIs.

openssl_encrypt versions before 1.4.0 contain a hardcoded default secret key in the standalone telemetry server configuration that is used for API key hashing. Attackers who know this default value can predict or forge API key hashes to compromise telemetry API authentication.

openssl_encrypt versions before 1.4.0 contain hardcoded database credentials in standalone server configuration files. Attackers on the same network can access PostgreSQL databases using well-known default credentials to retrieve sensitive data.

openssl_encrypt versions before 1.4.0 contain an authentication bypass vulnerability in CamelliaCipher that disables HMAC tag generation and verification when the PYTEST_CURRENT_TEST environment variable is set. Attackers with code execution can set this environment variable to produce unauthenticat

openssl_encrypt versions before 1.4.0 use HKDF with no salt and static info parameter in key normalization functions, reducing entropy extraction and determinism. Attackers can exploit predictable key derivation with identical inputs to weaken cryptographic security against multi-target attacks.

openssl_encrypt versions before 1.4.0 use a non-standard PBKDF2 key derivation construction with iterations=1 per call in an outer loop, creating a KDF whose security properties have not been formally analyzed. Attackers can exploit this weakened key derivation to more efficiently crack passwords pr

openssl_encrypt before 1.4.0 imports Python's non-cryptographic 'random' module (Mersenne Twister PRNG) at line 15 of openssl_encrypt/modules/pqc.py. No direct calls to random.* were present in the code, so no cryptographic operation is currently affected; however, the import creates a hazard that f

openssl_encrypt versions before 1.4.0 contain a plugin sandbox bypass vulnerability where the PluginImportGuard blocks a different set of modules than the AST analyzer's DANGEROUS_MODULES set. Attackers can bypass AST analysis through string obfuscation or encoding to import unblocked dangerous modu

openssl_encrypt versions before 1.4.0 contain a logging bug in restore_hidden_modules() that logs module counts after clearing, always showing zero restored modules and corrupting audit trails. Additionally, a race condition exists between module hiding and import hook installation where another thr

openssl_encrypt versions before 1.4.0 contain a path traversal vulnerability in the _is_safe_path method where the plugin_id parameter is not sanitized before constructing the plugin config directory path. Attackers can declare a malicious plugin_id containing path traversal sequences like '../' to

openssl_encrypt versions before 1.4.0 contain a sandbox bypass vulnerability where the plugin sandbox fails to restrict alternative file access methods like pathlib.Path and io.open. Attackers can import pathlib or io modules to read and write arbitrary files, completely bypassing the restricted_ope

openssl_encrypt versions before 1.4.0 contain an insecure default configuration that trusts the entire RFC 1918 private address space in IntegrityProxyConfig trusted_proxies. Attackers on private networks can forge client certificate headers to bypass mTLS authentication when ProxyAuth validation is

openssl_encrypt versions before 1.4.6 contain a key derivation flaw in sequential XOR composition mode where the last stage cancels out during key generation. When configured with a single KDF and no prior hashing stage, attackers can bypass memory-hard key derivation and perform offline password cr

openssl_encrypt (pip) versions <= 1.4.7 contain an information exposure vulnerability where the 'hsm fido2-test' and 'hsm onlykey-test' diagnostic commands unconditionally print the full derived hardware pepper as hex to stdout/stderr (crypt_cli.py, handle_hsm_command). The printed value can persist

ContextForge: DNS TOCTOU race condition causes SSRF protection bypass (`/admin/gateways/test`)

vLLM: Completion prompt lists fan out into unbounded engine requests

vLLM: Cross-User Data Leak Vulnerability

vLLM: Incomplete CVE-2025-62164 remediation can be bypassed by concurrent prompt parts

vLLM: ReDoS via structured_outputs.regex in the lm-format-enforcer backend (no compile timeout) — missed sibling of GHSA-rwxx-mrjm-wc2m

vLLM: Unauthenticated Internal Path and Username Disclosure via Validation Error Messages

atomic-agents-stack: Dashboard HTTP server path traversal allows arbitrary file read

Ansible FreeBSD Jail Connection Plugin: Jail escape via symlink following in put_file (host-side root mv)

MCP-for-Stata: Stata Command Injection via Unsanitized `package` in `ado_package_install`

Material for MkDocs: DOM XSS in search suggestions via query parameter

Apache Airflow missing team context permits cross-team Dag actions and XCom reads

Apache Airflow Task SDK fails to mask list-shaped JSON Variables

Apache Airflow bulk endpoints log Variable and Connection secrets in cleartext

Apache Airflow Backfill API parser discrepancy permits cross-DAG authorization bypass

Apache Airflow environment-variable secrets backend permits cross-team credential use

Apache Airflow Task SDK Callback deserialization can import arbitrary modules in the scheduler

Apache Airflow scheduler unsafely deserializes human-in-the-loop task next_kwargs

Apache Airflow Config API exposes team-scoped sensitive configuration values