Critical CVEs
New CVEs as they are published, with CVSS score, CISA KEV status, public exploits and the vendor's fix. Worth tracking, but only a small part of what Panop does: mapping what you expose to the internet, testing what an attacker can actually exploit, ranking it by business impact, and helping your team understand and fix it.
Feed at a glance
- Tracked
- 13360
- Critical
- 2486
- CISA KEV
- 125
- Public PoC
- 437
Most common types
- Broken Access Control 1691
- Remote Code Execution 1609
- Privilege Escalation 1307
- Authentication Bypass 1121
- Denial of Service 799
Published per day, last 14 days
Source: public CVE records (NVD, CISA KEV) · last published 3 Nov 2026
Critical CVEs
Newest first · severity as published by the vendor or NVD · page 149 of 166
- 9.1 /10 CriticalCVE-2026-57145
PraisonAI - Path Traversal
PraisonAI < 4.6.62 contains a path traversal vulnerability caused by lack of traversal rejection and symlink resolution in file operations in multiedit.py, letting prompt-influenced agents read or overwrite files, exploit requires LLM-controlled filepath input.
- 9.8 /10 CriticalCVE-2026-57131 No auth
PraisonAI - Broken Access Control
PraisonAI < 4.6.58 contains a broken access control caused by missing authentication and per-job authorization in praisonai.jobs.server.create_app mounting praisonai.jobs.router.create_router under /api/v1/runs, letting network attackers list, read, cancel, delete jobs, and execute unauthorized agents, exploit requires network access.
- 9.8 /10 CriticalCVE-2026-57127
PraisonAI - Authentication Bypass
PraisonAI < 4.6.58 contains an authentication bypass caused by APIKeyAuthMiddleware or JWTAuthMiddleware forwarding requests without required keys, letting unauthenticated clients access recipe execution and connected tools, exploit requires missing environment keys.
- 9.8 /10 CriticalCVE-2026-57125 No auth
PraisonAI - Command Injection
PraisonAI < 4.6.59 and praisonaiagents < 1.6.59 contain a command injection caused by unauthenticated POST /api/v1/runs accepting attacker-controlled agent_yaml and bypassing approval checks, letting remote attackers execute arbitrary OS commands without credentials or interaction.
- 9.8 /10 CriticalCVE-2026-57124
PraisonAI - Command Injection
PraisonAI < 4.6.59 contains a command injection caused by unauthenticated POST /api/mcp/connect accepting caller-controlled command and args, letting unauthenticated remote attackers execute commands as the UI service account, exploit requires network access to the UI host.
- 9.8 /10 CriticalCVE-2026-57123 No auth
PraisonAI praisonaiagents - Broken Access Control
PraisonAI praisonaiagents < 1.6.59 contains a broken access control caused by missing authentication and origin-validation in ToolsMCPServer routes, letting any reachable client invoke registered tools, exploit requires network access.
- 9.8 /10 CriticalCVE-2026-55209 No auth
resdata - Buffer Overflow
resdata < 6.2.9 contains a buffer overflow and out-of-bounds read caused by insufficient validation of numeric fields and array indexes in GRDECL file parsing, letting remote attackers cause memory corruption or denial of service, exploit requires untrusted GRDECL file input.
- 9.8 /10 CriticalCVE-2026-54334 No auth
UEFI Firmware Parser - Buffer Overflow
UEFI Firmware Parser < 1.14 contains a buffer overflow caused by improper bounds checking in ReadCLen() during parsing of Tiano or EFI compressed firmware, letting attackers cause heap corruption, crash, or potential code execution, exploit requires crafted firmware input.
- 9.8 /10 CriticalCVE-2026-54333 No auth
UEFI Firmware Parser - Buffer Overflow
UEFI Firmware Parser < 1.14 contains a buffer overflow caused by lack of validation of bit-length values in MakeTable() during Tiano or EFI compressed firmware parsing, letting attackers cause stack corruption and potential code execution, exploit requires crafted firmware bitstream.
- 9.1 /10 CriticalCVE-2026-53713 Auth required
Envoy Gateway - Path Traversal
Envoy Gateway < 1.7.4 and < 1.8.1 contains a file system read caused by improper path normalization in to_absolute_normalized_path in internal/gatewayapi/luavalidator/security.lua, letting attackers read arbitrary files from the gateway controller pod, exploit requires submission of Lua code through EnvoyExtensionPolicy.
- 9.1 /10 CriticalCVE-2026-50006 No auth
Anyquery - SQL Injection
Anyquery < 0.4.5 contains an unrestricted ATTACH DATABASE vulnerability in its MySQL-compatible server port, letting remote attackers create or overwrite arbitrary files, causing filesystem integrity loss and denial of service, exploit requires network access.
- 9.1 /10 CriticalCVE-2026-43790 No auth
Apple macOS - Denial of Service
Apple macOS Golden Gate 27, macOS Sequoia 15.8, and macOS Tahoe 26.7 contain a memory corruption vulnerability caused by improper memory handling, letting remote attackers cause system termination or kernel memory corruption, exploit requires network access.
- 9.8 /10 CriticalCVE-2026-20353
CVE-2026-20353
As part of Cisco's ongoing commitment to proactive security and product quality, the Cisco Secure Email Gateway and Cisco Secure Email and Web Manager engineering team has conducted a comprehensive internal security review. This review resulted in software hardening releases that address multiple internally discovered vulnerabilities. The vulnerabilities tracked by CVE-2026-20353 are related to issues with improper control of a resource through its lifetime that are grouped under the Common Weakness Enumeration (CWE) Pillar CWE-664.
- 9.0 /10 CriticalCVE-2026-18119 Auth required
Concrete CMS - Stored XSS & Privilege Escalation
Concrete CMS < 9.5.3 contains a stored cross-site scripting caused by unsanitized custom style values in the Block Design dialog, letting editor-level users execute scripts in administrator sessions and escalate privileges, exploit requires editor access.
- 9.9 /10 CriticalCVE-2026-16338 Auth required
IBM DataStage - Unrestricted File Upload
IBM DataStage on Cloud Pak for Data 5.4.0.0 contains an unrestricted file write vulnerability caused by improper validation of file paths, letting remote authenticated attackers write arbitrary files.