RuntimeDefault is the container runtime’s built-in seccomp profile. It provides a standard baseline of syscall restrictions without requiring a custom policy for each workload. It is often a practical starting point, but it may be either too permissive or too restrictive for applications with unusual syscall behavior.
What RuntimeDefault does
RuntimeDefault is the container runtime’s built-in seccomp profile. It gives a baseline syscall filter that most workloads can use without custom tuning, which makes it a practical default for reducing attack surface. Its main value is that it starts from a safer-than-unrestricted posture while still letting many ordinary containers run normally.
Why syscall filtering matters
Seccomp works by limiting the system calls a process can make into the kernel. That matters because containers share the host kernel, so a process that can invoke a broader set of syscalls has more opportunities to probe the kernel, escape application constraints, or exercise kernel features the application never needed. A default profile such as RuntimeDefault is therefore a control for containment, not a complete security boundary.
Because the profile is generic, it reflects common container behaviour rather than the exact needs of one workload. That means it can reduce exposure for many services, but it can also block legitimate syscall patterns in specialised applications, especially where low-level networking, tracing, storage, or language runtime behaviour is unusual.
How to use RuntimeDefault effectively
RuntimeDefault is usually the right first step when an organisation wants to harden containers quickly and consistently. It works best when combined with workload testing, because the practical question is not whether the profile is “secure” in the abstract, but whether it is compatible with the specific process behaviour of the application being deployed.
In practice, teams treat it as a baseline for most services and then move to a tighter custom profile only where application evidence justifies it. That approach avoids the two common failure modes: leaving containers effectively unconfined, or shipping a custom profile that is so restrictive it breaks availability and pushes operators to disable the control entirely.
Compatibility, tuning, and rollout
Not every workload fits a generic profile cleanly. A process may fail only under certain code paths, kernel features, or library calls, which makes seccomp issues look intermittent and hard to diagnose. When that happens, the operational goal is to identify which syscall family is required and whether the application can be redesigned, isolated, or granted only the narrow exception it genuinely needs.
Because RuntimeDefault is runtime-provided, behaviour can vary across container platforms and distributions. Teams should therefore validate it in the same environment where the workload will actually run, rather than assuming a profile that works in development will behave identically in production.
Risk and Threat Considerations
Container seccomp baselines reduce kernel exposure, but they do not eliminate risk. If RuntimeDefault is too permissive for a sensitive workload, an attacker who gains code execution may retain enough syscall access to support exploitation, privilege escalation, or kernel attack surface probing. If it is too restrictive, the usual failure is not direct compromise but operational workarounds that weaken the control or force exceptions.
Failure mechanism: Excessively broad syscall permissions leave more kernel paths available to malicious or unexpected code, while over-restriction can trigger exceptions, fallback configurations, or unsafe disablement of seccomp.
Impact: The likely outcomes are larger exploit surface, weaker containment, and avoidable reliability problems that erode confidence in the control.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
CIS Controls v8 and NIST SP 800-53 Rev 5 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| CIS Controls v8 | CIS-4 — Secure Configuration of Enterprise Assets and Software | Seccomp profiles are part of secure container and host configuration. |
| Recommendation — Apply CIS-4 to standardize RuntimeDefault-like hardening baselines and review exceptions for container workloads. | ||
| NIST SP 800-53 Rev 5 | SC-7 — Boundary Protection | Syscall filtering constrains process-to-kernel boundaries and limits reachable attack paths. |
| CM-7 — Least Functionality | RuntimeDefault implements a least-functionality baseline by denying unneeded syscalls. | |
| SI-4 — System Monitoring | Seccomp policy changes and blocked syscalls are security-relevant events to monitor. | |
| Recommendation — Use SC-7 to restrict container kernel exposure and block unnecessary syscall paths. Use CM-7 to remove unnecessary syscall capability and permit only required runtime behavior. Use SI-4 to alert on blocked syscall patterns, profile drift, and exception-driven weakening. | ||
Practitioner Guidance
What to watch for: Validate RuntimeDefault against real application behaviour, not just against deployment success. A profile is working only if it blocks unnecessary syscalls without causing repeated operational exceptions or brittle overrides.
Governance implication: Treat seccomp profile choice as an application security decision, not a platform checkbox. Owners should know which workloads rely on the default profile, which need custom policy, and where exceptions are approved so the baseline stays intentional rather than accidental.
Related resources from NHI Mgmt Group
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on October 30, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org