agentleFS
EvilFreelancerEvilFreelancer/secas/.agents/skills/container-security/SKILL.md

container-security

Skill11 starsChanged 3 months agoApache-2.0View on GitHub

Container and Kubernetes security assessment — image vulnerability scanning, SBOM diff analysis, K8s cluster auditing, RBAC privilege mapping, NetworkPolicy review, container escape testing, and runtime monitoring (Falco/Tetragon). Use when scanning Docker/OCI images, auditing K8s clusters, reviewing Dockerfiles, diffing SBOMs across releases, analyzing RBAC, or assessing container runtime posture. Triggers on requests involving Trivy, Grype, Syft, Kubescape, kube-bench, Falco, container escapes, or CIS Docker/K8s benchmarks.

  • Installs packages

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run concurrently (no shared state, no rate conflicts):

    line 70
  2. Very specific:

    All findings MUST conform to schemas/finding.json.

    line 126

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Container Security

Thin router for container and Kubernetes security assessments. Load the reference, workflow, or payload file you need — do not read all of them.

When to Use

  • Scanning Docker/OCI images for CVEs (single image or pipeline)
  • Generating or attesting SBOMs (Syft / CycloneDX / SPDX)
  • Diffing SBOMs across releases to flag newly-introduced CVEs
  • Auditing a live Kubernetes cluster (CIS / NSA / MITRE)
  • Mapping K8s RBAC and finding privilege-escalation paths
  • Reviewing NetworkPolicy coverage
  • Authoring Falco / Tetragon runtime rules
  • Analyzing or testing container escape vectors (authorized only)
  • Docker daemon / host hardening review

Trigger Phrases

"scan this image", "trivy / grype / syft", "audit k8s cluster", "kube-bench", "kubescape", "run CIS benchmark", "check RBAC", "networkpolicy coverage", "falco rule", "container escape", "SBOM diff", "new CVEs since last release".

When NOT to Use This Skill

RequestUse instead
Static scan of K8s YAML / Helm / Kustomize before deployiac-security
Terraform / CloudFormation / Pulumi misconfigiac-security
EKS / GKE / AKS control-plane or managed-service misconfigexploiting-cloud-platforms
Cloud IAM misconfiguration (beyond K8s RBAC)exploiting-cloud-platforms
Application-code vulns inside the containerauditing-code-for-vulnerabilities
Third-party library CVEs at source-code levelauditing-supply-chain
API endpoints exposed by containerized servicestesting-apis

Rule of thumb: pre-deployment YAML → iac-security; running cluster or built image → this skill.

Decision Tree

Target?
|
|-- Built/registry image ----------> workflows/image_scan.md
|     \-- two images to compare? --> workflows/sbom_diff.md    (FLAGSHIP)
|
|-- Dockerfile source -------------> examples/vulnerable_dockerfile.md
|                                    + hadolint (references/image_scanning.md)
|
|-- Live K8s cluster --------------> workflows/cluster_audit.md
|     |-- RBAC deep-dive ----------> workflows/rbac_analysis.md
|     |-- Network policy gap ------> workflows/network_policy_review.md
|     \-- CIS benchmark only ------> references/kubernetes_hardening.md
|
|-- Runtime monitoring ------------> references/runtime_security.md
|                                    + examples/falco_custom_rule.yaml
|
\-- Escape testing (authorized) --> references/container_escape.md
                                     + payloads/container_escape_poc.md

Parallelism Hints

Run concurrently (no shared state, no rate conflicts):

  • Trivy + Grype + Syft + Hadolint on the same image — always parallel
  • Scanners across multiple image tags — one sub-agent per tag
  • kube-bench across nodes — deploy as DaemonSet, collect in parallel
  • Kubescape frameworks (nsa, mitre, cis) — parallel invocations
  • RBAC object fetches (clusterroles, roles, bindings) — parallel
  • SBOM diff Grype scans (prior vs current) — parallel
  • Per-principal RBAC graph expansion — parallel (stateless)

Sequential only:

  • Cosign attestation (needs SBOM first)
  • Cross-scanner CVE consensus merge (needs all scanner outputs)
  • Finding emission (needs evaluated data)

Sub-Agent Delegation

  • Multi-image pipeline: spawn one sub-agent per image; each runs workflows/image_scan.md and returns finding records. Parent merges.
  • Multi-cluster environment: one sub-agent per cluster; each runs workflows/cluster_audit.md with its own kubecontext.
  • Cluster audit fan-out: parent runs inventory, then spawns sub-agents for rbac_analysis, network_policy_review, and runtime log triage in parallel.
  • SBOM diff per service: one sub-agent per service in a monorepo release; parent aggregates the "newly-introduced CVEs" roll-up.

Reasoning Budget

TaskBudget
Running a scanner and capturing outputMinimal
Applying CIS checks, filling checklistsMinimal
SBOM canonicalization and package diffMinimal
Cross-scanner CVE consensusModerate
RBAC privilege-graph traversal (principal -> cluster-admin)Extended
Container escape chain compositionExtended
SBOM diff CVE exposure scoring + waiver decisionExtended
Runtime alert triage (chaining Falco events to intent)Moderate

Extended thinking pays off where the answer requires composing many small facts into an attack path or risk decision; it does not help when running a scanner and reporting output verbatim.

Multimodal Hooks

  • Image layer histograms — attach Syft/Dive output screenshots to evidence for bloat/secret-layer findings.
  • RBAC graphs — rbac-tool viz produces graphviz; render as PNG and embed in report for stakeholder clarity.
  • Falco/Tetragon alert screenshots — useful for runtime-detection evidence bundles.

Structured Output

All findings MUST conform to schemas/finding.json.

Container-security-specific fields: affected.image_digest, affected.image_tag, affected.cluster_name, affected.namespace, affected.resource_kind, affected.service_account, cve, cvss, fixed_version, cis_control, and the sbom_diff block for diff findings.

Workflow Index

WorkflowPurpose
workflows/image_scan.mdBuild-time / registry image vuln scan with consensus merge
workflows/sbom_diff.mdFlagship. Compare prior vs current SBOM; flag new CVEs
workflows/cluster_audit.mdLive K8s cluster assessment (CIS / NSA / MITRE)
workflows/rbac_analysis.mdRBAC privilege mapping and attack-path search
workflows/network_policy_review.mdNetworkPolicy coverage + enforcement validation

Payloads Index

PayloadPurpose
payloads/container_escape_poc.mdDocumented escape PoCs (authorized testing only)

References Index

ReferencePurpose
references/image_scanning.mdTrivy, Grype, Syft, Clair, Snyk, Hadolint commands
references/kubernetes_hardening.mdkube-bench, Kubescape, CIS K8s mapping, PSS
references/container_escape.mdEscape vectors, capabilities, runtime CVEs
references/runtime_security.mdFalco, Tetragon, eBPF, rule authoring
references/bounty_patterns_2024_2026.mdPost-2023 bounty TTPs (CVE-2024-21626 runC, CVE-2025-23266 NVIDIA, SA token theft, RoleBinding privesc)

Examples Index

ExamplePurpose
examples/falco_custom_rule.yamlReady-to-load Falco rule pack
examples/vulnerable_dockerfile.mdCommon Dockerfile anti-patterns + fixes

Templates Index

TemplatePurpose
templates/assessment_report_template.mdEnd-of-engagement deliverable skeleton

Tools

ToolPurposeInstall
TrivyImage / FS vuln + SBOM + IaCbrew install trivy
GrypeImage / SBOM vulnbrew install grype
SyftSBOM generatorbrew install syft
HadolintDockerfile lintbrew install hadolint
KubescapeK8s security platformcurl -s https://raw.githubusercontent.com/kubescape/kubescape/master/install.sh | /bin/bash
kube-benchCIS K8s benchmarkbrew install kube-bench
kube-hunterK8s pentest reconpip install kube-hunter
FalcoRuntime security (eBPF)Helm chart falcosecurity/falco
TetragoneBPF detection + enforcementHelm chart cilium/tetragon
Docker BenchDocker CIS benchmarkgit clone https://github.com/docker/docker-bench-security.git
CosignSign / attest images + SBOMsbrew install cosign
CraneRegistry ops (resolve digest)brew install crane
rbac-toolRBAC graph visualizationbrew install insights-engineering/tap/rbac-tool

Last Validated

2026-04. Minimum tool versions: Trivy 0.59, Grype 0.87, Syft 1.18, Kubescape 3.0, kube-bench 0.10 (CIS K8s v1.9), Falco 0.38, Tetragon 1.2, Hadolint 2.12, Cosign 2.4.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.