CRITICAL🇵🇱 Wersja polska

CVE-2026-44336

CVSS 9.4v4.0pub. 2026-05-08upd. 2026-05-11

PraisonAI is a multi-agent teams system. Prior to version 4.6.34, PraisonAI's MCP (Model Context Protocol) server (praisonai mcp serve) registers four file-handling tools by default — praisonai.rules.create, praisonai.rules.show, praisonai.rules.delete, and praisonai.workflow.show. Each accepts a path or filename string from MCP tools/call arguments and joins it onto ~/.praison/rules/ (or, for workflow.show, accepts an absolute path) with no containment check. The JSON-RPC dispatcher passes params["arguments"] blind to each handler via **kwargs without validating against the advertised input schema. By setting rule_name="../../<some-path>" an attacker walks out of the rules directory and writes any file the running user can write. Dropping a Python .pth file into the user site-packages directory escalates this primitive to arbitrary code execution in any subsequent Python process the user spawns — the next praisonai CLI invocation, an IDE script run, the user's python REPL, or any background Python service. This issue has been patched in version 4.6.34.

🤖 AI Analysis
How it works

The MCP server (`praisonai mcp serve`) registers four file handling tools that accept a path or filename parameter from JSON-RPC call arguments and combine it directly with the `~/.praison/rules/` directory without any containment check. The JSON-RPC dispatcher passes the `params["arguments"]` parameters directly to each handler without validation against the declared input schema. An attacker can set the value `rule_name="../../<any-path>"`, which allows escaping the rules directory and writing a file to any location accessible to the current user. Placing a malicious `.pth` file in the user's `site-packages` directory results in execution of arbitrary Python code every time a Python process is subsequently launched — including when invoking the `praisonai` CLI, running a script in an IDE, a REPL session, or a background-running Python service.

Impact

An attacker can write arbitrary files within the scope of permissions of the running user, and by writing a malicious `.pth` file to the `site-packages` directory, can gain persistent and automatic execution of arbitrary code in all subsequent Python processes of that user.

Mitigation & patch

PraisonAI must be updated to version 4.6.34 or later, in which the vendor introduced appropriate path validation mechanisms. Details are available in the vendor's references: https://github.com/MervinPraison/PraisonAI/security/advisories/GHSA-9mqq-jqxf-grvw

Who is affected

PraisonAI (Praison product) in versions prior to 4.6.34 that use the MCP server functionality (`praisonai mcp serve`).

Analysis generated by Claude AI (Anthropic) based on NVD data. Always verify with vendor.
CVSS Vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
  • Praison Praisonai

    APP
    Praison
    < 4.6.34
🔵
CHECK WITH VENDOR
No clear patch data available. Check vendor references.
Tags
RCEPath Traversal
CWE
References

Related vulnerabilities

CVE-2026-41497CRITICAL9.8PL ✓same product

Command Injection w PraisonAI — brak walidacji poleceń MCP

CVE-2026-40288CRITICAL9.8PL ✓same product

PraisonAI — RCE i command injection przez niezaufowane pliki YAML

CVE-2026-40313CRITICAL9.1PL ✓same product

PraisonAI – wyciek tokenów GitHub przez atak ArtiPACKED w CI/CD

CVE-2026-40289CRITICAL9.1PL ✓same product

PraisonAI – nieuwierzytelnione przejęcie sesji przeglądarki przez WebSocket

CVE-2026-40157CRITICAL9.4PL ✓same product

Path traversal w PraisonAI — nadpisywanie plików przez złośliwy pakiet .praison