a2a-security-review
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Review a Kubernetes cluster's security across control plane, RBAC, workload configuration, network policy, secrets, and admission control, mapped to the CIS Kubernetes Benchmark and the 4Cs model. Use for a comprehensive cluster security assessment of a cluster you're authorized
$ npx -y skills add jassics/awesome-claude-security --skill k8s-cluster-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/k8s-cluster-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
Review a Kubernetes cluster's security across control plane, RBAC, workload configuration, network policy, secrets, and admission control, mapped to the CIS Kubernetes Benchmark and the 4Cs model. Use for a comprehensive cluster security assessment of a cluster you're authorized
name: k8s-cluster-review description: >- Review a Kubernetes cluster's security across control plane, RBAC, workload configuration, network policy, secrets, and admission control, mapped to the CIS Kubernetes Benchmark and the 4Cs model. Use for a comprehensive cluster security assessment of a cluster you're authorized to review.
A cluster assessment across all major control areas, each with findings, severity, and remediation, anchored to the CIS Kubernetes Benchmark.
1. **Control plane** — API server flags, anonymous/authn-authz config, etcd encryption & access, kubelet config, audit logging. (Managed clusters: provider owns some — cross-ref `cloud-security`.) 2. **RBAC & identity** — least privilege, cluster-admin sprawl, dangerous verbs, service-account token use. (Deep dive: `k8s-rbac-review`.) 3. **Workloads** — Pod Security Standards/admission, privileged/hostPath/hostNetwork pods, securityContext, capabilities. (Deep dive: `k8s-workload-hardening`.) 4. **Network** — default-deny NetworkPolicies, namespace isolation, exposed services/ LoadBalancers, ingress. 5. **Secrets** — etcd encryption at rest, secret access scope, secrets in env/images, external secret managers. 6. **Admission & supply chain** — admission controllers/policy engines (e.g. OPA/ Kyverno), image provenance/signing, registry trust, image scanning.
1. Establish scope and read access (`kubectl`/manifests/IaC). Note managed vs. self-managed (who owns the control plane). 2. Walk each area with `reference.md`; delegate RBAC and workload depth to the companion skills. 3. Record: area · control (CIS ref) · finding · severity · remediation. 4. Score (`security-reporting:cvss`) and rank.
A cluster report grouped by area with a CIS-mapped findings table + ranked top risks. Confirmed issues → `security-reporting:finding`.
Think in **4Cs** (Cloud, Cluster, Container, Code): a hardened pod on a wide-open API server is still exposed. On managed clusters, confirm which controls the provider owns vs. you. Default-deny network policy and restricting privileged pods are the highest-leverage wins.
A Claude Code plugin marketplace for the full cybersecurity & GenAI-security lifecycle — from recon and threat modeling to detection engineering, GRC, and CISO-level strategy. A pentester knows which OWASP test bends a broken-access-control endpoint.
Repo: jassics/awesome-claude-security
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Test the agent execution harness/runtime itself — LangChain/LangGraph, AutoGen, CrewAI, custom ReAct-style loops, or computer-use/browser-use agents — for…
Assess an autonomous / tool-using AI agent for security end-to-end: tool privileges, autonomy and approval boundaries, excessive agency, memory/state…
Test what an AI agent will actually do without human confirmation, including under injected-goal / prompt-injection scenarios, to validate its autonomy and…
Review the security of MCP (Model Context Protocol) servers/clients an agent uses: server trust tier, tool/resource description and result poisoning,…
Inventory the tools/functions an AI agent can call and audit their privileges, side effects, and approval requirements to find excessive-agency and…