Conciseness

The body is dominated by executable commands with tight, value-adding comments (capability hex masks, exact field selectors), but includes a few explanations of concepts Claude already knows, e.g. 'hostPID lets you see all node processes' and the nsenter explanation. It fits the score-4 anchor (efficient, minor instances of over-explanation that could be trimmed); not 5 because a handful of clarifying sentences restate common knowledge.

Actionability

Fully copy-paste-ready commands throughout: enumeration one-liners, mount/chroot escape, docker/ctr socket abuse, a complete cgroup release_agent script, a malicious Dockerfile, and a full kubectl --overrides privileged-pod manifest. This matches the score-5 anchor (specific examples cover the common cases); score 4 would understate the breadth of ready-to-run coverage.

Workflow Clarity

Clear Phase 1 → Phase 2 → Phase 3 sequence with per-primitive detection steps and an explicit decision rule ('check each, escalate via the first that works'). It is not 5 because there is no post-escape verification step confirming success (e.g. checking hostname/`id` against the host) and no validate-retry loop; it is above 3 because the detection checks act as real checkpoints before escalation.

Progressive Disclosure

Well-organized single-file skill with clear section headers (enumeration, per-technique subsections, post-escape, OPSEC, references) and no nested or buried references. It fits the score-4 anchor (good structure, content appropriately placed, minor organization gaps — e.g. the runC CVE-2024-21626 detail could live in a separate reference file); not 5 because no one-level-deep reference split exists for the more specialized CVE material.

Specificity

The description lists multiple specific concrete techniques — 'privileged container abuse, hostPath mount escape, hostPID/hostIPC, capability misuse (SYS_ADMIN, SYS_PTRACE), runC CVE chains' — comprehensively covering the escape primitives. It matches the score-5 anchor listing multiple specific concrete actions with comprehensive coverage, and is not the level below (4) because coverage is not merely 'several' actions but spans the full technique set the skill body addresses.

Completeness

The 'what' is explicit and detailed, and 'Pivots from RCE-in-pod to full node compromise' states a concrete precondition that functions as a when-clause ('when you have RCE inside a pod'). It is not 5 because the trigger guidance is not phrased as an explicit 'Use when...' clause with trigger phrases; it is above 3 because 'when' is more than weakly implied — the starting condition is named.

Trigger Term Quality

It contains good natural keywords a user would say ('kubernetes pod escape', 'privileged', 'hostPath', 'runC'), but common variations like 'k8s', 'container escape', and 'container break out / breakout' are absent from the description itself. Not 5 because synonym coverage is incomplete; not 3 because the terms present go well beyond one or two generic domain words.

Distinctiveness Conflict Risk

A clear niche (Kubernetes pod escape to node) with distinct triggers like 'hostPath mount escape' and 'runC CVE chains' that no neighboring skill would claim. Minimal conflict risk, matching the score-5 anchor; score 4 ('mostly distinct') would understate how unmistakable this trigger set is.