a2a-security-review
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Audit Kubernetes RBAC for least privilege and privilege-escalation paths — cluster-admin sprawl, wildcard/dangerous verbs, risky bindings, and service-account token exposure. Use to review who can do what in a cluster and find escalation to cluster-admin.
$ npx -y skills add jassics/awesome-claude-security --skill k8s-rbac-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/k8s-rbac-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
Audit Kubernetes RBAC for least privilege and privilege-escalation paths — cluster-admin sprawl, wildcard/dangerous verbs, risky bindings, and service-account token exposure. Use to review who can do what in a cluster and find escalation to cluster-admin.
name: k8s-rbac-review description: >- Audit Kubernetes RBAC for least privilege and privilege-escalation paths — cluster-admin sprawl, wildcard/dangerous verbs, risky bindings, and service-account token exposure. Use to review who can do what in a cluster and find escalation to cluster-admin.
An RBAC risk assessment: which subjects (users, groups, service accounts) hold excessive or dangerous permissions, and which paths lead to cluster-admin — with least-privilege fixes.
`rolebindings`/`clusterrolebindings`; `escalate`/`bind` verbs; `impersonate`.
`pods/exec`, `pods/attach`, `create` on `pods/ephemeralcontainers`.
cluster-admin via token theft).
cluster-admin bindings.
1. Enumerate Roles/ClusterRoles and their bindings; map subject → effective permissions (consider tooling like `kubectl auth can-i`, rbac audit tools). 2. Flag wildcard/admin grants, the dangerous verbs above, and broad secret access. 3. Trace escalation: can a namespace-scoped subject reach cluster-admin (e.g. via pod creation onto a privileged node, secret theft, or role self-grant)? Model the worst path (`security-diagramming:attack-tree`). 4. Record: subject · grant · risk · escalation? · least-priv recommendation.
An RBAC findings table + an escalation-path diagram for the worst case. Confirmed issues → `security-reporting:finding` (rate escalation-to-cluster-admin high+).
In Kubernetes, broad `secrets` read or `pods/exec` plus a privileged node is effectively cluster-admin — treat those as escalation, not "just read access." Recommend namespaced, least-privilege Roles and disabling unneeded SA token auto-mount.
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…