CVE-2026-47398

PRAISONAI vulnerability CVE-2026-47398: two unguarded calls to spec.loader.exec_module in praisonai/agents_generator.py allow arbitrary code execution when module_path from YAML is loaded. The issue exists in master prior to 4.6.40; fix is to gate by PRAISONAI_ALLOW_LOCAL_TOOLS and validate modul...

PraisonAI AgentOS - Information Disclosure

PraisonAI's AgentOS FastAPI application server exposes an unauthenticated GET /api/agents endpoint that lists every registered agent's name, role and the opening of its instructions system prompt. No authentication is enforced on the route, allowing a remote attacker to enumerate agent...

PYSEC-2026-2946 PraisonAI: Coarse-Grained Tool Approval Cache Bypasses Per-Invocation Consent for Shell Commands

Summary The approval system in PraisonAI Agents caches tool approval decisions by tool name only, not by invocation arguments. Once a user approves executecommand for any command e.g., ls -la, all subsequent executecommand calls in that execution context bypass the approval prompt entirely...

CVE-2026-40148

PraisonAI is a multi-agent teams system. Prior to 4.5.128, the safeextractall function in PraisonAI's recipe registry validates archive members against path traversal attacks but performs no checks on individual member sizes, cumulative extracted size, or member count before calling tar.extractal...

EUVD-2026-20638

PraisonAI Vulnerable to Remote Code Execution via YAML Deserialization in Agent Definition Loading...

CVE-2026-34935

PraisonAI is a multi-agent teams system. From version 4.5.15 to before version 4.5.69, the --mcp CLI argument is passed directly to shlex.split and forwarded through the call chain to anyio.openprocess with no validation, allowlist check, or sanitization at any hop, allowing arbitrary OS command...