// Security Advisory Triage

PraisonAI 1.7.2 and earlier expose agent endpoints without authentication

GitHub advisories — npm · published 2026-10-09 06:13 UTC
ACTION NEEDED PATCH_NOW SECURITYAPI
WHAT CHANGEDGitHub published CVE-2026-61426 on 8 October 2026. PraisonAI's AgentOS binds to the network without authentication by default, allowing unauthenticated callers to invoke agents and read their instructions. The vulnerability is rated high severity. Version 1.7.3 patches it.
WHY IT MATTERSAny deployment running the default configuration exposes agent logic and prompts to anyone who can reach the service. An attacker can execute arbitrary agent workflows and extract the instruction sets that define agent behavior.
WHAT TO DOUpgrade praisonai to 1.7.3. If you cannot upgrade immediately, verify whether your deployment binds to 0.0.0.0 or a public interface and restrict access at the network layer until you patch.
Model-reported confidence 95% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-6wjp-v33h-5cvq
+summary: PraisonAI: AgentOS defaults to network-exposed no-auth mode, allowing unauthenticated agent invocation and instruction disclosure
+severity: high
+published_at: 2026-10-08T22:01:29Z
+cve_id: CVE-2026-61426
+vulnerabilities[].package.name: praisonai
+vulnerabilities[].first_patched_version: 1.7.3
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.92, needed 0.5
triage reasoningAI agent framework with default no-auth network exposure and instruction disclosure affects services running AI agents.
signalsAI provider SDK/framework affected (PraisonAI agent framework) · unauthenticated access vulnerability in default configuration · instruction disclosure risk (prompt/agent instruction leakage) · high severity rating · patched version available (1.7.3) · new advisory entry with CVE assignment
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)