Research demo
IdentityMesh: agentic lateral movement through merged identities across connected systems
AIR-2025-0068 AIR record · 2025-07-lasso-identitymesh
From the supporting source
The IdentityMesh vulnerability exploits a fundamental architectural weakness in Agentic AI systems: the single source of privileges problem.lasso.security
The record
Disclosed
2025-07-30
Vendor
multiple
Product
MCP-based agents, AI browsers (incl. Perplexity Comet), custom crew.ai agents
Risk domain
Access control and permissions › Privilege escalation
Also touches
Data exfiltration channels, Confused deputy
System type
Generative model
Who acted
Agent acted — the agent moved laterally across connected systems on merged privileges
Attack surface
Multi-agent system
Autonomy level
Semi-autonomous
Initial vector
Indirect prompt injection connected system
Impact type
Lateral movement, Data exfiltration, Privilege abuse
Confidence
High
Submitted by
EAI Internal
Crosswalk
- OWASP Agentic Top 10 (2026)
- ASI03 – Identity and Privilege Abuse
- MITRE ATLAS
- AML.T0055 – Unsecured Credentials
- NIST AI RMF
- GOVERN 6.1
- EU AI Act
- Article 13
- ISO/IEC
- TR 24028; 42001; 23894
- AIUC-1
- B. Security
- Scope
- tool misuse
- Mitigation
- Apply least-privilege principle; Regular policy audits.
- Taxonomy
- Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
Enterprise assistants integrating multiple systems under a single agent identity
Capability abused
Taxonomy & references
Curator notes
Authors Bar Lanyado, Ophir Dror, Or Oxenberg. Three attack scenarios demonstrated; PoC research, no in-the-wild exploitation claimed. Some coverage dates this to March 2026; the post itself is dated 2025-07-30.
Supporting sources
Cite as
Enkrypt AI. "IdentityMesh: agentic lateral movement through merged identities across connected systems". AIR, AIR-2025-0068. https://agentincident.vercel.app/i/AIR-2025-0068/ (accessed 2026-10-19).