policy-as-code – Coding skill for AI agents
Implement policy as code with OPA, Sentinel, and Kyverno. Automate policy enforcement in CI/CD and infrastructure. Use when enforcing compliance through automation.
Install the policy-as-code skill with one command
Pick the tool you use, copy the command and paste it in your terminal. The skill lands in that tool's skills folder and is ready in your next session.
Claude Code
mkdir -p ~/.claude/skills && curl -fsSL https://skill.dn1nguoi.com/z/policy-as-code-41c51.tgz | tar xz -C ~/.claude/skills
Call it in Claude Code: /policy-as-code – then describe the task you want done.
Codex
mkdir -p ~/.codex/skills && curl -fsSL https://skill.dn1nguoi.com/z/policy-as-code-41c51.tgz | tar xz -C ~/.codex/skills
Call it in Codex: $policy-as-code – then describe the task you want done.
Hermes Agent
mkdir -p ~/.hermes/skills && curl -fsSL https://skill.dn1nguoi.com/z/policy-as-code-41c51.tgz | tar xz -C ~/.hermes/skills
Call it in Hermes Agent: /policy-as-code – then describe the task you want done.
DeepSeek
mkdir -p ~/.deepcode/skills && curl -fsSL https://skill.dn1nguoi.com/z/policy-as-code-41c51.tgz | tar xz -C ~/.deepcode/skills
Call it in DeepSeek: /policy-as-code – then describe the task you want done.
SKILL.md content
Policy as Code
Automate policy enforcement through code using OPA/Rego, Kyverno, Checkov, and CI/CD integration to prevent compliance violations before they reach production.
Enforcement contract
Examples cover selected controls, not full compliance. Terraform plan JSON can
contain secrets: restrict artifact permissions/retention and never upload it as a
public artifact. Planning runs providers/data sources and needs approved account
scope; do not run untrusted PR configuration with production credentials.
opa eval normally exits successfully even when deny contains findings; use a
reviewed aggregate decision or Conftest with all namespaces and fail on parse,
missing input, unknown values and missing required policy bundles. Resolve the
provider version/schema first: ACLs, encryption and tags use different resources.
The S3 versioning resource check cannot detect a bucket with no versioning resource;
add a graph relationship check. Kyverno examples require compatible CRDs, audit
rollout, full container/init/ephemeral coverage and separate digest/signature
verification: banning latest does not make another mutable tag trustworthy.
Open Policy Agent (OPA) Rego Policies
# deny_public_s3.rego - Deny S3 buckets with public access
package terraform.aws.s3
import rego.v1
deny contains msg if {
resource := input.resource_changes[_]
resource.type in {"aws_s3_bucket", "aws_s3_bucket_acl"}
resource.change.after.acl == "public-read"
msg := sprintf(
"S3 bucket '%s' has public-read ACL. All buckets must be private. [Policy: no-public-s3]",
[resource.address]
)
}
deny contains msg if {
resource := input.resource_changes[_]
resource.type in {"aws_s3_bucket", "aws_s3_bucket_acl"}
resource.change.after.acl == "public-read-write"
msg := sprintf(
"S3 bucket '%s' has public-read-write ACL. This is strictly prohibited. [Policy: no-public-s3]",
[resource.address]
)
}# require_encryption.rego - Require encryption on data stores
package terraform.aws.encryption
import rego.v1
deny contains msg if {
resource := input.resource_changes[_]
resource.type == "aws_db_instance"
resource.change.after != null
resource.change.after.storage_encrypted != true
msg := sprintf(
"RDS instance '%s' does not have storage encryption enabled. [Policy: require-rds-encryption]",
[resource.address]
)
}
deny contains msg if {
resource := input.resource_changes[_]
resource.type == "aws_ebs_volume"
resource.change.after != null
resource.change.after.encrypted != true
msg := sprintf(
"EBS volume '%s' is not encrypted. [Policy: require-ebs-encryption]",
[resource.address]
)
}
# S3 encryption is modeled by a separate Terraform resource in current providers.
# Join aws_s3_bucket_server_side_encryption_configuration to its bucket and inspect
# key/SSE settings; this snippet intentionally makes no S3 coverage claim.# require_tags.rego - Enforce mandatory tagging
package terraform.aws.tags
import rego.v1
required_tags := {"Environment", "Owner", "CostCenter", "DataClassification"}
deny contains msg if {
resource := input.resource_changes[_]
resource.mode == "managed"
resource.type in {"aws_instance", "aws_db_instance", "aws_ebs_volume", "aws_s3_bucket"}
resource.change.after != null
tags := object.get(resource.change.after, "tags_all", {})
missing := required_tags - {key | value := tags[key]; is_string(value); trim_space(value) != ""}
count(missing) > 0
msg := sprintf(
"Resource '%s' is missing required tags: %v. [Policy: required-tags]",
[resource.address, missing]
)
}Region enforcement must resolve each resource's provider alias/module mapping and
unknown expressions. Reading only provider_config.aws silently misses aliased or
variable-driven regions. Require a normalized resolved region per resource from a
reviewed adapter, and deny unknown mappings; no generic region policy is supplied.
# Evaluate OPA policies against Terraform plan
terraform plan -out=tfplan
terraform show -json tfplan > tfplan.json
# Run OPA evaluation
opa eval \
--data policies/ \
--input tfplan.json \
"data.terraform.aws.s3.deny" \
--format pretty
# Use conftest for easier CI integration
conftest test tfplan.json --policy policies/ --all-namespaces --output tableKyverno Kubernetes Policies
# require-labels.yaml - Enforce required labels on all pods
apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
name: require-labels
annotations:
policies.kyverno.io/title: Require Labels
policies.kyverno.io/category: Best Practices
policies.kyverno.io/severity: medium
spec:
validationFailureAction: Enforce
background: true
rules:
- name: check-required-labels
match:
any:
- resources:
kinds:
- Pod
validate:
message: >-
Labels 'app.kubernetes.io/name', 'app.kubernetes.io/version',
and 'team' are required on all Pods.
pattern:
metadata:
labels:
app.kubernetes.io/name: "?*"
app.kubernetes.io/version: "?*"
team: "?*"
---
# disallow-privileged.yaml - Block privileged containers
apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
name: disallow-privileged-containers
annotations:
policies.kyverno.io/title: Disallow Privileged Containers
policies.kyverno.io/category: Pod Security
policies.kyverno.io/severity: high
spec:
validationFailureAction: Enforce
background: true
rules:
- name: deny-privileged
match:
any:
- resources:
kinds:
- Pod
validate:
message: "Privileged containers are not allowed."
pattern:
spec:
containers:
- =(securityContext):
=(privileged): false
=(initContainers):
- =(securityContext):
=(privileged): false
=(ephemeralContainers):
- =(securityContext):
=(privileged): false
---
# require-resource-limits.yaml - Enforce resource limits
apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
name: require-resource-limits
annotations:
policies.kyverno.io/title: Require Resource Limits
policies.kyverno.io/severity: medium
spec:
validationFailureAction: Enforce
background: true
rules:
- name: check-resource-limits
match:
any:
- resources:
kinds:
- Pod
validate:
message: "All containers must have CPU and memory limits defined."
pattern:
spec:
containers:
- resources:
limits:
memory: "?*"
cpu: "?*"
---
# disallow-latest-tag.yaml - Block usage of 'latest' image tag
apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
name: disallow-latest-tag
annotations:
policies.kyverno.io/title: Disallow Latest Tag
policies.kyverno.io/severity: medium
spec:
validationFailureAction: Enforce
background: true
rules:
- name: validate-image-tag
match:
any:
- resources:
kinds:
- Pod
validate:
message: "Images must use a specific tag, not 'latest'."
pattern:
spec:
containers:
- image: "!*:latest & *:*"
---
# restrict-image-registries.yaml - Allow only approved registries
apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
name: restrict-image-registries
annotations:
policies.kyverno.io/title: Restrict Image Registries
policies.kyverno.io/severity: high
spec:
validationFailureAction: Enforce
background: true
rules:
- name: validate-registries
match:
any:
- resources:
kinds:
- Pod
validate:
message: >-
Images must come from approved registries:
123456789012.dkr.ecr.us-east-1.amazonaws.com or ghcr.io/your-org.
pattern:
spec:
containers:
- image: "123456789012.dkr.ecr.*.amazonaws.com/* | ghcr.io/your-org/*"
---
# NetworkPolicy presence cannot be checked by comparing a namespace string to itself.
# Use a reviewed apiCall context with scoped RBAC to list policies, then verify an
# applicable default-deny policy and CNI enforcement. This requires integration tests;
# absence of a real policy query must never be represented as a passing control.Checkov Custom Checks
# custom_checks/require_s3_versioning.py
from checkov.terraform.checks.resource.base_resource_check import BaseResourceCheck
from checkov.common.models.enums import CheckResult, CheckCategories
class S3Versioning(BaseResourceCheck):
def __init__(self):
name = "Ensure S3 bucket has versioning enabled"
id = "CUSTOM_S3_001"
supported_resources = ["aws_s3_bucket_versioning"]
categories = [CheckCategories.BACKUP_AND_RECOVERY]
sup… (full content is in the download package)
Related skills
- policy-library – Policy register: name, version, category, owner and approver, applies to, compliance…
- power-user-cultivation – When the user wants to identify and nurture developer advocates, build champion…
- pre-release-review – Run a read-only pre-release review for deploy readiness, migrations, config, secrets…
- privacy-mask – Mask, redact, anonymize and censor sensitive information (PII) in screenshots and images…
- probation-tracker – Probation register: employee, start and end dates, 30, 60 and 90-day review scores…
- project-based-performance – Project performance review: role, project and manager, delivery, quality and…
- push-skill-to-github – Commit and push skill changes to the configured skills repository after review and…
- pydantic-models-py – Create Pydantic models following the multi-model pattern for clean API contracts.