InfoResearchPreprintLLM-specific
Beyond the Patch: Reconstructing an Incomplete-Fix Patch Bypass in LLM Agent Frameworks with Control-Gated, Benign Probes
- Published
- Record updated
Summary
This pilot study tests whether patches for LLM agent framework vulnerabilities hold beyond the original trigger. Using benign, control-gated probes, the authors compare original-trigger replay with a sibling-path probe across PraisonAI, Langroid, Agno and CrewAI. At PraisonAI 4.6.31 the original tool-loading trigger is blocked, but the tool_override path still violates the invariant, reproducing the bypass of CVE-2026-40287 via CVE-2026-44334.
Mitigation
Neither tested path shows a violation at PraisonAI 4.6.32. For CrewAI, a violation was observed in tested release 1.11.0 and confinement in 1.11.1, without claiming the earliest fixed release.
Topics
Related items
- InfoTrust-gated robust median: Root-free defense against backdoor attacks in Non-IID federated learningSimilar attack · Elsevier Security Journals
- HighCVE-2026-67455: NVIDIA Model Optimizer for Linux and Windows deserialization of untrusted dataSimilar attack · NVD/CVE Database
- MediumThe risks of device code authentication for AI coding agentsSimilar attack · Microsoft Security Blog
- MediumGrok Bots Attempt to Access Corporate Data Without User ApprovalSimilar attack · Dark Reading
- HighSiemens Mendix Studio ProSimilar attack · CISA Cybersecurity Advisories