Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add yindia/rootcause --skill k8s-autoscalinggit clone --depth 1 https://github.com/yindia/rootcauseWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/yindia/rootcause/k8s-autoscaling)<a href="https://agentmods.dev/skills/yindia/rootcause/k8s-autoscaling"><img src="https://agentmods.dev/badge/skills/yindia/rootcause/k8s-autoscaling/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yindia/rootcause/k8s-autoscaling"><img src="https://agentmods.dev/badge/skills/yindia/rootcause/k8s-autoscaling.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.02244 |
| Opus 5 | $0.00000 | $0.01122 |
| Sonnet 5 | $0.00000 | $0.00449 |
| Haiku 4.5 | $0.00000 | $0.00224 |
Grade A, and why
k8s-autoscaling scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 369 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: k8s-autoscaling
Deep autoscaling diagnostics and tuning for three scaling paths:
- Horizontal Pod Autoscaler (HPA),
- Vertical Pod Autoscaler (VPA),
- Karpenter node provisioning.
Use this skill to explain why scaling did or did not happen, then provide a cost-aware remediation path.
Trigger Phrases
Use this skill when the user mentions:
- hpa not scaling
- replicas stuck
- scale up too slow
- scale down never happens
- vpa recommendation needed
- oomkilled and right sizing
- pods pending no nodes
- karpenter not provisioning
- nodepool constraints
- nodeclass misconfiguration
- spot interruptions
- autoscaling cost too high
RootCause Tools Allowed
Only use these tool names in this skill:
k8s.hpa_debugk8s.vpa_debugk8s.resource_usagek8s.scale(requiresconfirm=true)k8s.describek8s.listkarpenter.statuskarpenter.node_provisioning_debugkarpenter.nodepool_debugkarpenter.nodeclass_debugkarpenter.interruption_debug
Autoscaling Model
Think in layers:
- Pod replica scaling is controlled by HPA.
- Pod request sizing is controlled by VPA.
- Node supply scaling is controlled by Karpenter.
A healthy system aligns all three layers.
Decision Tree
| Primary Symptom | Start Here | Next Branch |
|---|---|---|
| CPU high, replicas unchanged | k8s.hpa_debug |
validate metrics and maxReplicas |
| Recurrent OOMKilled | k8s.vpa_debug |
compare requests vs recommendations |
| Pods pending with scheduling errors | karpenter.node_provisioning_debug |
inspect NodePool/NodeClass |
| Node count spikes and costs jump | k8s.resource_usage |
check over-requesting and HPA sensitivity |
| Frequent node terminations | karpenter.interruption_debug |
check spot/interruption behavior |
Workflow A: HPA Diagnostics (CPU/Memory/Custom Metrics)
Step A1: Enumerate HPAs and targets
Use k8s.list:
namespace: checkout
resources:
- kind: HorizontalPodAutoscaler
- kind: Deployment
Confirm target references are valid.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 369 lines · 0 tokens per session scan A d3f1b59eead5
k8s-autoscaling is a skill published in the GitHub repository yindia/rootcause (42 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,244 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
bazel-k8s-expert
Expert knowledge for deploying Quarkus/Java applications to Kubernetes using Bazel. Covers rulesk8s, Helm, Kustomize, ConfigMaps, Secrets, and health probes.
kubernetes-review
Kubernetes & Cloud-Native Review: Reviews Kubernetes manifests, Helm charts, and cloud-native configurations for security, reliability, resource management, and best practices. Covers pods, deployments, services, ingress, RBAC, network policies, HPA, PDB, security contexts, and GitOps patterns. Use when the user…
kubernetes-operator
Deploy and manage applications on Kubernetes. Covers deployments, services, ingress, HPA, secrets, and production-grade cluster configuration.
securing-helm-chart-deployments
Secure Helm chart deployments by validating chart integrity, scanning templates for misconfigurations, and enforcing security contexts in Kubernetes releases.
k8s-network-policy
Review Kubernetes NetworkPolicies — what is actually permitted to reach a workload, and whether the answer can be trusted. Use when asked what can talk to a pod, whether a namespace is restricted, why traffic is being blocked, or for any security review of cluster network segmentation.
k8s-service-path
Trace the Kubernetes service path — Service to selector to pods to EndpointSlices to readiness, plus Ingress routing. Use when a service is getting no traffic, an ingress is not routing, or someone asks why a workload is unreachable inside a cluster.