Ruflo MCP Flaw Lets Unauthenticated Attackers Run Commands and Poison AI Memory
Summary: Researchers at Noma Labs found a CVSS 10.0 flaw, CVE-2026-59726, in Ruflo, an open-source agent meta-harness for Claude Code and Codex with more than 66,500 GitHub stars, affecting all versions before 3.16.3. The default docker-compose file bound port 3001 to 0.0.0.0, exposing an unauthenticated MCP bridge with 233 tools, so a single HTTP POST to terminal_execute gave a shell, and the maintainer shipped a fix within 24 hours of the June 30, 2026 disclosure.
Bottom line: An attacker with that shell could read every LLM provider API key in the container, harvest all stored conversations, and write a poisoned pattern into the AgentDB learning store that steers future AI outputs long after the intrusion ends - which is why Noma says a patch alone is not remediation and the keys and memory must be treated as compromised.
Read the original at The Hacker News ↗
- Topics
- Vulnerabilities & Patching
- Mentions
- CVE-2026-59726
- Related
- Attackers Exploit UnrealIRCd Flaw That Enables Server-Level Command Injection Cyber Risks · Nov 14, 2026U.S. CISA adds Check Point and UnrealIRCd flaws to its Known Exploited Vulnerabilities catalog Cyber Risks · Nov 14, 2026PostgreSQL Patches 52 Flaws, Including 22-Year-Old Bug Enabling Authenticated RCE Cyber Risks · Nov 13, 2026