karpenter-autoscaling

Karpenter for intelligent Kubernetes node autoscaling on EKS. Use when configuring node provisioning, optimizing costs with Spot instances, replacing Cluster Autoscaler, implementing consolidation, or achieving 20-70% cost savings.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

75%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 highly actionable, well-structured skill with extensive executable examples and clean reference signaling. Its main weakness is conciseness: marketing prose, a redundant quick-reference recap, repeated savings claims, and un-quarantined version tokens inflate the body without adding capability Claude lacks.

Suggestions

Cut the 'Real-World Results' and 'Cost Savings Formula' marketing recap and the redundant Quick Reference section; the savings range is already stated in the description and Overview.

Move time-sensitive version pins ('v1.0+', '1.0.0', '2025') into a short 'Version / compatibility' note so they don't bloat the evergreen procedural sections.

Push the full Spot-mix, consolidation-budget, and GPU NodePool YAML blocks into references/nodepools.md and reference.md and keep only one canonical minimal NodePool inline.

Dimension	Reasoning	Score
Conciseness	The bulk is lean actionable code, but the Overview/Real-World Results prose ('20% overall AWS bill reduction', '70% reduction in monthly compute costs') plus a redundant Quick Reference section that repeats the Helm install and basic NodePool, repeated savings figures throughout, and un-quarantined time-sensitive tokens ('v1.0', '2025', '1.0.0') add padding that could be trimmed, matching the 'mostly efficient but could be tightened' anchor rather than a higher one.	3 / 5
Actionability	Copy-paste-ready, executable guidance is pervasive — Helm install, kubectl apply/scale/logs, full NodePool and EC2NodeClass YAML, Terraform HCL, PDB and GPU pod manifests — and examples span the common cases (install, Spot mix, consolidation, migration, GPU, monitoring), matching the fully-executable top anchor.	5 / 5
Workflow Clarity	Quick Start and Workflows 1–5 are numbered with explicit verify steps ('kubectl get pods ...', 'Verify nodes', 'Verify installation') and the migration workflow includes a testing checklist plus a rollback plan; however most workflows lack an explicit validate→fix→retry gating loop, leaving minor validation gaps below the 5 anchor.	4 / 5
Progressive Disclosure	The body signals three one-level-deep, real references with bold 'See:' links (installation.md, nodepools.md, optimization.md, all present) and navigation is easy, but the ~800-line body inlines substantial detailed config (e.g., the full Spot/consolidation/GPU NodePool blocks) that could live in those references, a minor organization gap versus the cleanly-split 5 anchor.	4 / 5
	Total	16 / 20 Passed

Description

87%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 strong, third-person description that pairs a clear 'what' with an explicit multi-trigger 'when' clause, using natural EKS/Karpenter vocabulary. Minor tightening of the opening phrasing and broader trigger synonyms would push it to the top anchor.

Suggestions

Lead with a concrete action verb (e.g., 'Provisions and consolidates EC2 nodes for EKS autoscaling') instead of the abstract 'Karpenter for intelligent...autoscaling'.

Add a couple of common natural synonyms users say ('scale up nodes', 'node groups', 'bin-packing') to round out trigger coverage.

Dimension	Reasoning	Score
Specificity	Quotes 'configuring node provisioning, optimizing costs with Spot instances, replacing Cluster Autoscaler, implementing consolidation, or achieving 20-70% cost savings' list several specific actions, but the opening 'Karpenter for intelligent Kubernetes node autoscaling on EKS' states the domain more abstractly than with concrete action verbs, leaving minor gaps versus the comprehensive 5 anchor.	4 / 5
Completeness	It explicitly answers 'what' ('Karpenter for intelligent Kubernetes node autoscaling on EKS') and 'when' with concrete trigger phrases ('Use when configuring node provisioning, optimizing costs with Spot instances, replacing Cluster Autoscaler...'), matching the anchor that requires both answered clearly with explicit triggers.	5 / 5
Trigger Term Quality	Natural terms a user would say are present — 'EKS', 'Spot instances', 'Cluster Autoscaler', 'consolidation', 'node provisioning' — giving good coverage, but a few common variants like 'scale up/down', 'node groups', or 'bin-packing' are missing, so it does not fully reach the comprehensive 5 anchor.	4 / 5
Distinctiveness Conflict Risk	The Karpenter-on-EKS niche with triggers like 'replacing Cluster Autoscaler' and 'Spot instances' is distinct and unlikely to fire for unrelated skills, but 'cost savings' and 'consolidation' are broad enough to carry minor overlap risk with general cost-optimization guidance, placing it just below the clear-niche 5 anchor.	4 / 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
skill_md_line_count	SKILL.md is long (812 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: fernandezbaptiste/Skrillz
Path: skills/karpenter-autoscaling/SKILL.md
Commit: 93ed392

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.