Stop, Continue, or Ignore are the three common outcomes used to control how a workflow reacts to a failed custom step. Stop halts the run immediately, Continue proceeds but marks the run failed at the end, and Ignore proceeds while treating the overall run as successful.
Expanded Definition
Stop, Continue, or Ignore is a workflow error-handling decision that determines how an automated run behaves when a custom step fails. In NHI and agentic systems, the choice is not just operational convenience; it is a governance control over whether an execution path should abort, degrade, or proceed despite a failed security or validation action. Definitions vary across vendors, but the core distinction is consistent: Stop enforces immediate halt, Continue allows completion while preserving failure state, and Ignore suppresses the failure from the run outcome. That makes the setting especially important when a step handles secret retrieval, token exchange, policy checks, or approval gates. It is closely related to exception handling in automation and to control reliability in frameworks such as the NIST Cybersecurity Framework 2.0, where resilience depends on predictable response to failures. The most common misapplication is using Ignore on a security-critical step, which occurs when teams prioritise pipeline continuity over the integrity of identity, secrets, or authorization checks.
Examples and Use Cases
Implementing Stop, Continue, or Ignore rigorously often introduces operational friction, requiring organisations to weigh availability and developer velocity against the risk of silently accepting a failed control.
- Stop is appropriate when an AI agent cannot validate a secret before calling a privileged tool, because proceeding would create unauthorised execution risk.
- Continue can be useful for non-blocking observability tasks, such as logging or metrics enrichment, where the run should finish but the failure must still be visible for review.
- Ignore may be acceptable for low-impact optional enrichment steps, but only when a failure has no bearing on authorization, data access, or downstream integrity.
- In NHI governance, teams often pair this setting with lifecycle controls described in the Ultimate Guide to NHIs so that failed rotation or validation steps cannot be buried by automation defaults.
- During API key rotation, a Stop outcome is often chosen if a post-rotation verification call fails, because leaving the system in an unknown state can extend exposure windows.
Industry usage is still evolving, especially in agentic orchestration platforms where the same labels may map to different runtime semantics. When teams compare these modes to identity assurance practices in the NIST Cybersecurity Framework 2.0, the key question is whether failure should block trust decisions or merely annotate them.
Why It Matters in NHI Security
This setting matters because NHI failures are often invisible until they produce a breach, an access outage, or a broken control chain. If a failed credential validation step is ignored, an agent may continue operating with stale or unverified access. If a rotation check is allowed to continue without stopping the run, a compromised secret can remain active longer than intended. NHIMG research shows that 79% of organisations have experienced secrets leaks, with 77% of those incidents causing tangible damage, which illustrates how quickly automation mistakes become security events when failure handling is too permissive. The same pattern appears in zero trust programs, where 90% of IT leaders say properly managing NHIs is essential for success, because trust decisions depend on reliable enforcement rather than best effort execution. A Stop, Continue, or Ignore choice therefore becomes part of the identity control plane, not just a developer preference. Organisations typically encounter the operational cost only after a failed rotation, failed approval, or failed token exchange has already been bypassed, at which point the term becomes operationally unavoidable to address.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 and OWASP Agentic AI Top 10 address the attack and risk surface, while NIST CSF 2.0, NIST Zero Trust (SP 800-207) and NIST AI RMF set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-05 | Failure handling can hide unsafe NHI execution paths and weak policy enforcement. |
| OWASP Agentic AI Top 10 | A-04 | Agent workflows need explicit failure semantics for tool calls and task completion. |
| NIST CSF 2.0 | PR.AC-1 | Access control outcomes depend on predictable handling of failed identity checks. |
| NIST Zero Trust (SP 800-207) | TA | Zero Trust requires continuous verification, which fails if steps can be silently ignored. |
| NIST AI RMF | GOV-4 | Governance of automated systems includes defined escalation and exception handling. |
Treat failed authentication or authorization checks as trust-breaking events, not optional warnings.
Related resources from NHI Mgmt Group
- What breaks when fraud controls stop at onboarding and ignore payout time?
- What breaks when fraud controls stop at checkout and ignore returns or refund abuse?
- How should organisations stop auto-sync from turning desktops into repositories of credentials?
- How should teams stop LLMjacking when NHI secrets leak?
