hunt-k8s

Hunt Kubernetes & Docker — API anonymous access, kubelet 10250 exec (SPDY/WebSocket, NOT plain POST) and the simpler /run primitive, etcd 2379 unauth, dashboard skip-login, RBAC misconfig, secret/SA-token abuse, docker.sock host escape, runc/container-escape (Leaky Vessels CVE-2024-21626), API-server-mediated nodes/proxy RCE, EphemeralContainers node-shell, bound/projected SA-token audience+expiry abuse, admission-controller bypass, Helm/Tiller remnants. Use when target runs containerized infra, exposes K8s ports (6443/10250/10255/2379/8443), or cloud metadata reveals K8s service accounts.

Quality

96%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

92%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a lean, highly actionable offensive playbook with strong sequencing and rigorous validation gates, assuming Claude's domain competence throughout. Its one weakness is progressive disclosure: all material lives in one large monolithic file with no external references to offload the deeper CVE and per-phase detail.

Suggestions

Split the per-phase command catalogs and CVE writeups into reference files (e.g. references/kubelet-10250.md, references/etcd-2379.md, references/cve-gates.md) and link them one level deep from a shorter overview, so SKILL.md stays a launchpad.

Extract the CVE/version-gate table (CVE-2018-1002105, CVE-2024-21626, CVE-2019-5736, CVE-2022-0492, CVE-2020-8558, CVE-2022-3294) into a dedicated references/cve-gates.md so version-sensitive detail is isolated and easier to maintain.

Move the 'False-Positive Killers' and 'Validation Checklist' into a references/confirmation-gate.md referenced from the top, keeping the inline confirmation-gate summary as a pointer.

Dimension	Reasoning	Score
Conciseness	The body is dense with executable commands and precise caveats and never explains domain basics Claude already knows (no 'Kubernetes is a container orchestrator'); its length is breadth of attack surface, all signal, fitting 'every token earns its place'.	3 / 3
Actionability	Copy-paste-ready curl/etcdctl/kubectl/kubeletctl commands carry real flags, paths, and payloads — e.g. the docker.sock privileged-container PoC, SelfSubjectRulesReview JSON bodies, and the etcd v3 /v3/kv/range query — the top 'fully executable, copy-paste ready' anchor.	3 / 3
Workflow Clarity	Phases 1–8 give a clear sequence, and validation is explicitly present for these destructive/offensive operations — the 'OOB / Confirmation Gate', per-finding 'Validation Checklist' checkboxes, and 'False-Positive Killers' supply the checkpoints and feedback loops the top anchor requires.	3 / 3
Progressive Disclosure	It is a ~290-line single file with no references/, scripts/, or assets/ bundle and no external links; though well-sectioned into phases, content that could be split (detailed CVE writeups, per-phase command catalogs) is inline, matching 'content that should be separate is inline'.	2 / 3
	Total	11 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is highly specific, trigger-rich, and complete, naming concrete attack primitives and giving explicit 'Use when' guidance scoped to exposed K8s ports and cloud-metadata pivots. Voice is third-person/imperative with no first- or second-person slips.

Dimension	Reasoning	Score
Specificity	Enumerates many concrete actions — 'API anonymous access', 'kubelet 10250 exec (SPDY/WebSocket, NOT plain POST) and the simpler /run primitive', 'etcd 2379 unauth', 'docker.sock host escape', 'EphemeralContainers node-shell' — matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	Answers both what (the full vector enumeration) and when via an explicit 'Use when target runs containerized infra, exposes K8s ports (6443/10250/10255/2379/8443), or cloud metadata reveals K8s service accounts.' clause, the top anchor.	3 / 3
Trigger Term Quality	Natural pentest-facing terms dominate — 'Kubernetes & Docker', 'K8s ports (6443/10250/10255/2379/8443)', 'containerized infra', 'cloud metadata reveals K8s service accounts', 'docker.sock' — the terms an assessor would actually say, satisfying 'good coverage of natural terms'.	3 / 3
Distinctiveness Conflict Risk	It carves a clear offensive K8s/Docker hunting niche with distinct port- and primitive-based triggers, making conflict with unrelated skills unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: elementalsouls/Claude-BugHunter
Path: skills/hunt-k8s/SKILL.md
Commit: 6b9c96e

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.