TI-2026-34961 Aliases CVE-2026-47391 GHSA-vg22-4gmj-prxw

PraisonAI's unauthenticated A2A official example can reach real LLM-driven `eval()` tool execution

HIGH 82CWE-306Published 2026-07-21 · Updated 2026-07-23

Key facts

Attack vector: Network
Privileges required: None
User interaction: None
Attack complexity: Low
Impact type: Security bypass, Code execution
Confidentiality · integrity · availability: High · High · High
Fix status: Fixed

Attack profile from CVSS 3.1 · CNA. Fix status from CNA and GitHub. Impact type from CWE-306 and CWE-95.

Summary

PraisonAI is a multi-agent teams system. Prior to version 4.6.40, PraisonAI's first-party A2A server example exposes an unauthenticated A2A JSON-RPC endpoint and registers a `calculate(expression)` tool implemented with Python `eval()`. The example also binds to `0.0.0.0`. A remote unauthenticated attacker can send `message/send` to `/a2a`; the request reaches `agent.chat()`, and a real LLM can invoke the registered `calculate` tool. In testing with `gemini/gemini-2.5-flash-lite`, this resulted in arbitrary Python execution in the server process, confirmed by creation of a marker file from an unauthenticated HTTP request. The issue affects deployments following the official A2A example or similar unauthenticated public A2A deployments with unsafe tools. The default unauthenticated A2A surface also exposes task history and task cancellation APIs, increasing confidentiality and integrity impact. Version 4.6.40 patches the issue.

Why it matters now

Exploitation Proof of concept CISA SSVC · not in CISA KEV

EPSS 0.01498 74th percentileScored 2026-11-07

Fix Fixed PraisonAI (MervinPraison): 4.6.40PraisonAI (PyPI): 4.6.40

SSVC decision points

CISA ADP · SSVC 2.0.3

Exploitation: Poc
Automatable: Yes
Technical impact: Total

The decision (Track, Attend, Act) also depends on your mission and well-being impact; see CISA's SSVC guide.

Affected products

Vendor	Product	Affected versions	Fixed in	Status	Source
MervinPraison	PraisonAI	before 4.6.40	4.6.40	Patch available	CNA
—	PraisonAI (PyPI)	before 4.6.40	4.6.40	Patch available	GHSA-vg22-4gmj-prxw

PraisonAI VendorMervinPraison Affected before 4.6.40 Fixed in 4.6.40 Patch available CNA
PraisonAI (PyPI) Vendor— Affected before 4.6.40 Fixed in 4.6.40 Patch available GHSA-vg22-4gmj-prxw

CPE names Threatist (inferred) · 1

Inferred from how NVD named this product before; not given by a source.

CPE	Versions	Role	Vendor
cpe:2.3:a:praison:praisonai::::::::	—	—	MervinPraison

cpe:2.3:a:praison:praisonai:*:*:*:*:*:*:*:* Versions — — ·MervinPraison

Remediation

Fix · CNA PraisonAI (MervinPraison) Fixed in 4.6.40
Fix · GHSA-vg22-4gmj-prxw PraisonAI (PyPI) Fixed in 4.6.40 Advisory →

Timeline

2026-10-02 Threatist record created from CVE-2026-47391
2026-10-03 Alias added: GHSA-vg22-4gmj-prxw

References

Other 4

TI-2026-79619 Missing authentication in PraisonAI AgentOS exposes agent chat CVSS 3.1CRITICAL 9.8
TI-2026-71236 Missing access control in PraisonAI AgentTeam exposes agent endpoints CVSS 3.1CRITICAL 9.8
TI-2026-44052 PraisonAI before 1.6.78 Remote Code Execution via CodeAgent CVSS 4.0CRITICAL 10.0
TI-2026-41123 praisonai-platform 0.1.4 still boots on the hardcoded JWT secret dev-secret-change-me (default-open production guard) CVSS 3.1CRITICAL 9.8
TI-2026-41122 praisonai-platform: default JWT signing secret 'dev-secret-change-me' enables token forgery CVSS 3.1CRITICAL 9.8
TI-2026-41118 PraisonAI MCPServer exposes unauthenticated HTTP tools/call CVSS 3.1CRITICAL 9.8

See all 177 →

Sources

This product uses the NVD API but is not endorsed or certified by the NVD. Sources and licences