k8s-pod-escape

Kubernetes pod escape to node — privileged container abuse, hostPath mount escape, hostPID/hostIPC, capability misuse (SYS_ADMIN, SYS_PTRACE), runC CVE chains. Pivots from RCE-in-pod to full node compromise.

Quality

78%

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

Fix and improve this skill with Tessl

tessl review fix ./packages/decepticon/decepticon/skills/standard/cloud/container/k8s-pod-escape/SKILL.md

Quality

Content

82%Weight 40%Scale 1-5

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

A tight, highly actionable offensive playbook with copy-paste commands and a clear phased workflow; its main limitation is the absence of bundle files or offloaded reference material, leaving it as a single monolithic document.

Suggestions

Add explicit validation checkpoints between phases (e.g. 'Confirm CapEff includes the target bit before attempting the escape; if absent, move to 2.x') to strengthen the feedback loop.

Split the longer technique writeups (cgroup release_agent, runC CVE chain) into a references/ file linked from the body to improve progressive disclosure and keep SKILL.md as an overview.

Note kernel/version prerequisites inline at each technique (e.g. 'kernel < 5.8' for cgroup release_agent) rather than burying them in comments, so preconditions are scannable.

Dimension	Reasoning	Score
Conciseness	The body is dense and command-driven with terse inline comments, but includes a few explanatory glosses (e.g. CapEff hex meanings, 'nsenter -t 1' rationale) that could be trimmed; it does not pad with concepts Claude already knows.	4 / 5
Actionability	Provides fully executable, copy-paste-ready bash for every major escape path (privileged mount, hostPath/docker.sock, hostPID+SYS_PTRACE, cgroup release_agent, runC CVE, SA-token API abuse) covering the common cases concretely.	5 / 5
Workflow Clarity	Clear Phase 1 (enumerate) → Phase 2 (detect primitives, 'escalate via the first that works') → Phase 3 (pivot) sequence with capability detection before action serving as validation; minor gaps in explicit verify-and-retry checkpoints keep it just below 5.	4 / 5
Progressive Disclosure	No bundle files exist, so the single ~150-line file is scored on its own structure; it is well-organized into phased sections with subsections, an OPSEC block, and a references list, though everything is inlined with no offloaded detail files.	4 / 5
	Total	17 / 20 Passed

Description

75%Weight 40%Scale 1-5

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

A specific, technically rich description that clearly conveys its niche and concrete capabilities, but it lacks an explicit 'Use when...' trigger clause and leans heavily on jargon over natural user phrasing.

Suggestions

Add an explicit trigger clause, e.g. 'Use when you have RCE inside a Kubernetes pod and need to escape to the host node or pivot to the cluster.'

Soften jargon with natural synonyms users actually say — 'container escape', 'break out of a pod', 'breakout' — alongside the CVE/capability terms.

Move the CVE numbers and capability bit-patterns out of the description into the body or metadata tags to keep the description readable as a trigger.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete techniques — 'privileged container abuse, hostPath mount escape, hostPID/hostIPC, capability misuse (SYS_ADMIN, SYS_PTRACE), runC CVE chains' and 'Pivots from RCE-in-pod to full node compromise' — giving comprehensive, specific coverage rather than vague language.	5 / 5
Completeness	It clearly states what the skill does, but there is no 'Use when...' clause or equivalent explicit trigger guidance; the 'when' is only weakly implied by 'Pivots from RCE-in-pod', so per the rubric cap completeness stays at 3.	3 / 5
Trigger Term Quality	Natural terms like 'pod escape', 'privileged container', and 'hostPath' are present, but the description is dominated by technical jargon (CVE numbers, capability names) and omits common synonyms such as 'container escape' or 'breakout'.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear, narrow niche (Kubernetes pod escape to node) with distinct triggers and minimal overlap risk against other skills.	5 / 5
	Total	17 / 20 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
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: PurpleAILAB/Decepticon
Path: packages/decepticon/decepticon/skills/standard/cloud/container/k8s-pod-escape/SKILL.md
Commit: 0cf691e

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.