Implement Kubernetes Pod Security Admission (PSA) with controlled policy rollout (`audit`, `warn`, `enforce`) to raise workload security posture without destabilizing operations.
Scanned 9/5/2026
Install to Claude Code
npx -y skills add devton/agentic-workflow-blueprint --skill implementing-pod-security-admission-controller --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Implementing Pod Security Admission Controller?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/devton-implementing-pod-security-admission-controller)More formats (shields.io, HTML) on the badges page.
## agentic-workflows-blueprint.workflow.implementing-pod-security-admission-controller
### Goal
Implement Kubernetes Pod Security Admission (PSA) with controlled policy rollout (`audit`, `warn`, `enforce`) to raise workload security posture without destabilizing operations.
### Scope
- Applies to: namespace-level Pod Security Standard enforcement and cluster-level PSA defaults.
- Does not cover: custom policy engines for rules beyond PSA baseline/restricted semantics.
### Triggers
- "Enable Pod Security Admission"
- "Migrate from PodSecurityPolicy to PSA"
- "Enforce baseline/restricted policies per namespace"
- "Audit pod security violations before enforcement"
### Inputs
- `clusterVersion`: Kubernetes version and API server config model
- `namespacePolicyMap`: desired profile per namespace (`privileged`, `baseline`, `restricted`)
- `enforcementStrategy`: rollout sequence and timing by environment
- `exemptionList` (optional): system/critical namespaces with justification
- `changeWindow` (optional): execution window for enforcement changes
- `rollbackPlan`: policy rollback and emergency exemption procedure
### Invariants
- Production namespaces should not jump directly to strict enforcement without audit evidence.
- Enforcement versions must be pinned for deterministic behavior across upgrades.
- Exemptions must be minimal, explicit, and justified.
- Any breaking enforcement change requires rollback readiness before rollout.
### Procedure
1. **Assess current workload posture**
- Inventory namespace workloads and identify likely PSA violations.
- If migrating from PSP, map existing controls to PSA profiles.
2. **Define target policy map**
- Assign namespace profiles: system namespaces (`privileged` when justified), staging (`baseline`), production (`restricted` target).
- Define pinned version labels for enforce/audit/warn.
3. **Start with audit and warn**
- Apply `audit` and `warn` labels before `enforce` in target namespaces.
- Collect warning/audit evidence and prioritize remediations.
4. **Remediate violating workloads**
- Update securityContext and pod specs to meet target profile.
- Validate restricted/baseline compliance for critical workloads.
5. **Enable enforcement progressively**
- Apply `enforce` per namespace in staged sequence aligned to `changeWindow`.
- Monitor admission rejections and service health after each stage.
6. **Configure cluster defaults (optional)**
- When required, configure API server admission control defaults and controlled exemptions.
- Validate control-plane behavior after config rollout.
7. **Verify and stabilize**
- Confirm pod creations match expected allow/deny behavior.
- Document residual exceptions and follow-up remediation actions.
### Outputs
- Namespace PSA label plan and applied label evidence.
- Violation/remediation report from audit and warn phases.
- Enforcement rollout report with accepted/rejected workload outcomes.
- Exemption register with owner, reason, and review date.
### Review gate
- [ ] Target policy profile per namespace is explicit and version-pinned.
- [ ] Audit/warn evidence exists before strict enforcement in sensitive namespaces.
- [ ] Violating workloads are remediated or formally excepted.
- [ ] Enforcement rollout is staged and operationally validated.
- [ ] Exemptions are justified, minimal, and tracked.
### References
- `../../SKILL.md`
- `../os-platform/SKILL.md`
- `../review/SKILL.md`
- [Interactive HTML View](./README.html)
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!