MervinPraison PraisonAI
Published Atomicorp research notes for CVEs affecting MervinPraison PraisonAI where WAF protections were observed during testing or engineering review.
These notes are not intended to represent the full set of vulnerabilities mitigated by Atomicorp products and should not be interpreted as a coverage matrix, certification list, or census of protected CVEs.
Atomicorp WAF protections are primarily designed around attack techniques and exploit behaviors rather than individual vulnerability identifiers. A published research note documents a positive research finding related to that CVE. If a CVE is absent from these notes, no conclusion should be drawn about protection status.
| CVE | Vulnerability | CVSS | Severity | Rules Observed |
|---|---|---|---|---|
| CVE-2026-47391 | PraisonAI's unauthenticated A2A official example can reach real LLM-driven eval() tool execution | 9.8 (v3.1) | Critical | 340014 , 340023 , 340029 , 344360 , 344361 , 344363 , 344364 , 344366 , 344370 , 393655 |
| CVE-2026-57124 | PraisonAI UI MCP connect endpoint allows unauthenticated local command execution | 9.8 (v3.1) | Critical | 340014 , 340023 , 340029 , 344360 , 344361 , 344363 , 344364 , 344366 , 344370 , 390904 , 393655 |
| CVE-2026-57131 | praisonai: Jobs API exposes agent-execution endpoints with no authentication | 9.8 (v3.1) | Critical | 340014 , 340023 , 340029 , 344360 , 344361 , 344363 , 344364 , 344366 , 344370 , 393655 |
| CVE-2026-56840 | PraisonAI: HTTPApproval dashboard renders tool arguments as raw HTML, allowing approval-page XSS to approve dangerous to | 8.8 (v3.1) | High | 333140 , 333141 , 340095 , 340147 , 340148 , 341256 , 342259 , 346755 , 350147 , 350148 |
| CVE-2026-47394 | PraisonAI vulnerable to unauthenticated arbitrary file read via MCP workflow.show, workflow.validate, deploy.validate | 8.7 (v4.0) | High | 340007 , 344360 , 390709 |
| CVE-2026-57126 | praisonaiagents: SSRF guard validates literal IPs only and never resolves DNS | 8.5 (v3.1) | High | 337109 , 337110 , 340165 , 344360 , 347009 , 390722 , 398021 , 398022 |
| CVE-2026-61429 | PraisonAI before 1.6.78 SSRF via Crawl4AI Chromium backend | 8.4 (v4.0) | High | 337109 , 337110 , 344360 , 398021 , 398022 |
| CVE-2026-61430 | PraisonAI before 1.6.78 DNS Rebinding SSRF via web_crawl | 8.4 (v4.0) | High | 337109 , 337110 , 344360 , 398021 , 398022 |
| CVE-2026-47398 | PraisonAI: Arbitrary code execution via unguarded spec.loader.exec_module in agents_generator.py - sibling of CVE-20 | 8.1 (v3.1) | High | 340014 , 340023 , 340029 , 344360 , 344361 , 344363 , 344364 , 344366 , 344370 , 393655 |
| CVE-2026-57119 | PraisonAI: Unauthenticated Local File Inclusion via agent_file path in the Jobs API | 7.5 (v3.1) | High | 340007 , 344360 , 390709 |
| CVE-2026-57114 | PraisonAI: Jobs webhook SSRF protection bypass via DNS rebinding | 7.2 (v3.1) | High | 337109 , 337110 , 344360 , 398021 , 398022 |
| CVE-2026-55537 | PraisonAI: Webhook SSRF via DNS fail-open in JobSubmitRequest.validate_webhook_url() — bypass of CVE-2026-40114 | 7.1 (v3.1) | High | 337109 , 337110 , 344360 , 398021 , 398022 |
| CVE-2026-60091 | PraisonAI before 4.6.78 Unauthenticated SSRF via webhook_url | 6.9 (v4.0) | Medium | 337109 , 337110 , 344360 , 398021 , 398022 |