Skip to main content
No isolation or failure boundary between agents — a failure in one propagates unchecked to all downstream

​
How to resolve

Add failure isolation boundaries between agents. Implement circuit breakers that trip when error rates exceed thresholds. Limit blast radius of individual agent failures.

​
Risk

Governance/Compliance This vulnerability falls under ASI08:2026 — Cascading Failures in the OWASP Top 10 for Agentic Applications. Errors in one agent propagate across multi-agent pipelines with compounding destructive impact. Security No isolation or failure boundary between agents — a failure in one propagates unchecked to all downstream. If exploited, this can compromise the agent’s integrity, confidentiality, or availability, potentially affecting downstream systems and data.

​
Explanation

No isolation or failure boundary between agents — a failure in one propagates unchecked to all downstream. This falls under ASI08 (Cascading Failures): Errors in one agent propagate across multi-agent pipelines with compounding destructive impact.

​
Specifications

Trigger
  • Agent scan detects no isolation or failure boundary between agents — a failure in one propagates unchecked to all downstream
The scanner looks for patterns such as:
  • no error handling between agent task calls
  • agent chain with no retry or fallback logic
  • downstream agent receives unvalidated upstream output
Severity
  • HIGH
Applies to: Agent