Monthly
Hash collisions in the inference request cache of NVIDIA Megatron-LM let an attacker who can already submit inference requests tamper with cached data and obtain partial information about other users' cached requests and responses, yielding a CVSS 7.1 impact (AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:N). Because the cache key relies on a weak hash (CWE-328), success depends on the attacker engineering colliding inputs against an enabled, shared inference cache, which is why the vector rates complexity as high rather than trivial. No public exploit code has been identified at time of analysis and the CVE is not listed in CISA KEV; this is a moderate-severity issue rather than a top-tier emergency, and the NVIDIA product-security advisory is the authoritative source for affected builds and fixed versions.
Token validation bypass in Apache CXF's STSTokenValidator and Security Token Service (STS) allows unauthenticated remote attackers to impersonate another user by presenting a forged UsernameToken or self-signed SAML Assertion whose 32-bit Java hash collides with an entry already held in the validated-token cache. Because a cache hit is treated as proof of prior validation, the forged token is accepted without password validation, signature trust verification, or any call to the STS, and can subsequently be exchanged through STS validation or renewal for STS-signed tokens issued for the victim identity. Exploitation is constrained in practice: the target deployment must use STSTokenValidator or STS token validation/renewal with the validated-token cache enabled, and a legitimately validated token for the identity being impersonated must already be present in that cache; no public exploit code has been identified at time of analysis, KEV status is negative, and EPSS is low (~0.17%, 6th percentile), yet the headline CVSS of 9.1 reflects high confidentiality and integrity impact if the preconditions are met. Vendor patches are available in 4.2.4, 4.1.9, and 3.6.13.
Credential disclosure in R-SOFT DMS lets an attacker who obtains the stored superadmin password hash recover the plaintext password, because credentials are protected with a non-salted, nested MD5 hash that is trivially reversible via brute-force and precomputed rainbow tables. The superadmin account is high-value and cannot be rotated through the UI - the password can only be changed by editing the configuration file - so recovered credentials remain valid indefinitely. There is no public exploit identified at time of analysis, and the flaw is not listed in CISA KEV.
Cache key collision in Angular's @angular/common HttpTransferCache allows remote attackers to poison Server-Side Rendering (SSR) cache entries and replace responses for sensitive endpoints with attacker-controlled content. The weak 32-bit DJB2-like polynomial rolling hash used for TransferState cache keys is trivially brute-forceable, enabling state poisoning, DOM-based XSS, and information leakage when a victim follows a crafted link. No public exploit identified at time of analysis, but the vulnerability was discovered and reported by Google DeepMind's CodeMender and is patched in Angular 22.0.1, 21.2.17, and 20.3.25.
Weak password hashing on the GNCC GP5 IP camera (firmware v7.1.76) allows attackers who obtain the password hash to recover the root credential via offline brute-force, yielding full device takeover. The CVSS 9.8 score reflects network-reachable impact, but real-world exploitation first requires extracting the hash from the device; no public exploit identified at time of analysis and EPSS is very low at 0.02%.
Algorithmic complexity attack in jq JSON processor allows remote denial of service via hash collision exploitation. An attacker can craft a ~100KB JSON object with precomputed colliding keys that degrade hash table performance from O(1) to O(n²), causing severe CPU exhaustion in unauthenticated network contexts including CI/CD pipelines and web services. The vulnerability stems from a hardcoded MurmurHash3 seed (0x432A9843) that enables offline collision calculation. Fixed in commit 0c7d133c3c7e37c00b6d46b658a02244fdd3c784. CVSS 7.5 (High severity, network-exploitable, no authentication required). No public exploit identified at time of analysis, but attack technique is well-documented and feasible.
Insufficient input padding in soroban-poseidon's Poseidon V1 hash function enables attackers to forge hash collisions by appending zeros to shorter inputs, allowing distinct messages to produce identical hashes when the input count is less than the sponge rate. This vulnerability affects any Soroban smart contract relying on PoseidonSponge or poseidon_hash for cryptographic integrity, potentially compromising authentication, signature verification, or other security mechanisms that depend on hash uniqueness. No patch is currently available.
Salt deduction in PostgreSQL Anonymizer prior to 3.2.3 allows an authenticated low-privileged masked role to bypass RESTRICTED function protections by calling anon.hash() from inside a sub-select, collecting (seed, hash_output) pairs and performing an offline brute-force attack to recover the masking salt. This is a low-severity information-disclosure issue (CVSS 4.3, PR:L) that weakens masking guarantees but does not directly reveal plaintext or grant privilege escalation; no public exploit identified at time of analysis and no CISA KEV entry was identified. The issue is fixed in PostgreSQL Anonymizer 3.2.3 and later.
Artifact Integrity Validation in wandb 0.25.2.dev1 uses MD5 (a cryptographically broken hash algorithm) within ArtifactManifestEntry.download in wandb/sdk/lib/hashutil.py, enabling a sufficiently resourced attacker with write access to artifact storage or a man-in-the-middle network position to substitute a malicious artifact file that shares the same MD5 digest as a legitimate one, bypassing download integrity checks. The CVSS 4.0 base score of 2.3 reflects high attack complexity and the requirement for authenticated access (PR:L), placing real-world exploitability firmly in the theoretical-to-low range. No public exploit exists and the vulnerability has no CISA KEV listing; an upstream fix via SHA-256 supplementation is available as an unmerged GitHub PR (#12031).
Weak hashing in LangGraph's Task Result Cache exposes low-privilege authenticated users to hash collision attacks that can leak cached task results across isolation boundaries. All LangGraph versions through 1.2.4 are affected via the `_freeze` function in `libs/langgraph/langgraph/_internal/_cache.py`. A public proof-of-concept is disclosed at GitHub issue #8009, though the CVSS 4.0 score of 2.3 and AC:H rating reflect genuinely high exploitation difficulty; this is not listed in CISA KEV and no active exploitation has been reported.
Hash collisions in the inference request cache of NVIDIA Megatron-LM let an attacker who can already submit inference requests tamper with cached data and obtain partial information about other users' cached requests and responses, yielding a CVSS 7.1 impact (AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:N). Because the cache key relies on a weak hash (CWE-328), success depends on the attacker engineering colliding inputs against an enabled, shared inference cache, which is why the vector rates complexity as high rather than trivial. No public exploit code has been identified at time of analysis and the CVE is not listed in CISA KEV; this is a moderate-severity issue rather than a top-tier emergency, and the NVIDIA product-security advisory is the authoritative source for affected builds and fixed versions.
Token validation bypass in Apache CXF's STSTokenValidator and Security Token Service (STS) allows unauthenticated remote attackers to impersonate another user by presenting a forged UsernameToken or self-signed SAML Assertion whose 32-bit Java hash collides with an entry already held in the validated-token cache. Because a cache hit is treated as proof of prior validation, the forged token is accepted without password validation, signature trust verification, or any call to the STS, and can subsequently be exchanged through STS validation or renewal for STS-signed tokens issued for the victim identity. Exploitation is constrained in practice: the target deployment must use STSTokenValidator or STS token validation/renewal with the validated-token cache enabled, and a legitimately validated token for the identity being impersonated must already be present in that cache; no public exploit code has been identified at time of analysis, KEV status is negative, and EPSS is low (~0.17%, 6th percentile), yet the headline CVSS of 9.1 reflects high confidentiality and integrity impact if the preconditions are met. Vendor patches are available in 4.2.4, 4.1.9, and 3.6.13.
Credential disclosure in R-SOFT DMS lets an attacker who obtains the stored superadmin password hash recover the plaintext password, because credentials are protected with a non-salted, nested MD5 hash that is trivially reversible via brute-force and precomputed rainbow tables. The superadmin account is high-value and cannot be rotated through the UI - the password can only be changed by editing the configuration file - so recovered credentials remain valid indefinitely. There is no public exploit identified at time of analysis, and the flaw is not listed in CISA KEV.
Cache key collision in Angular's @angular/common HttpTransferCache allows remote attackers to poison Server-Side Rendering (SSR) cache entries and replace responses for sensitive endpoints with attacker-controlled content. The weak 32-bit DJB2-like polynomial rolling hash used for TransferState cache keys is trivially brute-forceable, enabling state poisoning, DOM-based XSS, and information leakage when a victim follows a crafted link. No public exploit identified at time of analysis, but the vulnerability was discovered and reported by Google DeepMind's CodeMender and is patched in Angular 22.0.1, 21.2.17, and 20.3.25.
Weak password hashing on the GNCC GP5 IP camera (firmware v7.1.76) allows attackers who obtain the password hash to recover the root credential via offline brute-force, yielding full device takeover. The CVSS 9.8 score reflects network-reachable impact, but real-world exploitation first requires extracting the hash from the device; no public exploit identified at time of analysis and EPSS is very low at 0.02%.
Algorithmic complexity attack in jq JSON processor allows remote denial of service via hash collision exploitation. An attacker can craft a ~100KB JSON object with precomputed colliding keys that degrade hash table performance from O(1) to O(n²), causing severe CPU exhaustion in unauthenticated network contexts including CI/CD pipelines and web services. The vulnerability stems from a hardcoded MurmurHash3 seed (0x432A9843) that enables offline collision calculation. Fixed in commit 0c7d133c3c7e37c00b6d46b658a02244fdd3c784. CVSS 7.5 (High severity, network-exploitable, no authentication required). No public exploit identified at time of analysis, but attack technique is well-documented and feasible.
Insufficient input padding in soroban-poseidon's Poseidon V1 hash function enables attackers to forge hash collisions by appending zeros to shorter inputs, allowing distinct messages to produce identical hashes when the input count is less than the sponge rate. This vulnerability affects any Soroban smart contract relying on PoseidonSponge or poseidon_hash for cryptographic integrity, potentially compromising authentication, signature verification, or other security mechanisms that depend on hash uniqueness. No patch is currently available.
Salt deduction in PostgreSQL Anonymizer prior to 3.2.3 allows an authenticated low-privileged masked role to bypass RESTRICTED function protections by calling anon.hash() from inside a sub-select, collecting (seed, hash_output) pairs and performing an offline brute-force attack to recover the masking salt. This is a low-severity information-disclosure issue (CVSS 4.3, PR:L) that weakens masking guarantees but does not directly reveal plaintext or grant privilege escalation; no public exploit identified at time of analysis and no CISA KEV entry was identified. The issue is fixed in PostgreSQL Anonymizer 3.2.3 and later.
Artifact Integrity Validation in wandb 0.25.2.dev1 uses MD5 (a cryptographically broken hash algorithm) within ArtifactManifestEntry.download in wandb/sdk/lib/hashutil.py, enabling a sufficiently resourced attacker with write access to artifact storage or a man-in-the-middle network position to substitute a malicious artifact file that shares the same MD5 digest as a legitimate one, bypassing download integrity checks. The CVSS 4.0 base score of 2.3 reflects high attack complexity and the requirement for authenticated access (PR:L), placing real-world exploitability firmly in the theoretical-to-low range. No public exploit exists and the vulnerability has no CISA KEV listing; an upstream fix via SHA-256 supplementation is available as an unmerged GitHub PR (#12031).
Weak hashing in LangGraph's Task Result Cache exposes low-privilege authenticated users to hash collision attacks that can leak cached task results across isolation boundaries. All LangGraph versions through 1.2.4 are affected via the `_freeze` function in `libs/langgraph/langgraph/_internal/_cache.py`. A public proof-of-concept is disclosed at GitHub issue #8009, though the CVSS 4.0 score of 2.3 and AC:H rating reflect genuinely high exploitation difficulty; this is not listed in CISA KEV and no active exploitation has been reported.