AGENT SKILL

k8s-client-js

暂无简介

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 13
Forks
2

03 SKILL.md 正文

/k8s-client-js

When to Use

  • Agent spawns, monitors, and terminates child agent Pods programmatically
  • Scale a Deployment based on task queue depth
  • Stream logs from a running agent pod back to the orchestrator
  • Exec a command inside a container for health checks or remediation

Do NOT use for

  • Infrastructure-as-code (use [[helm-chart-packaging]] or [[argocd-gitops]])
  • High-frequency operations (K8s API has rate limits ~100 req/s)

Setup and authentication

import * as k8s from '@kubernetes/client-node'

const kc = new k8s.KubeConfig()

// Auto-detect: in-cluster (ServiceAccount) or ~/.kube/config
kc.loadFromDefault()

const coreApi  = kc.makeApiClient(k8s.CoreV1Api)
const appsApi  = kc.makeApiClient(k8s.AppsV1Api)
const batchApi = kc.makeApiClient(k8s.BatchV1Api)

Spawn an agent Pod

async function spawnAgentPod(agentId: string, tier: string): Promise<string> {
  const pod = await coreApi.createNamespacedPod('yamtam', {
    metadata: {
      name:   `agent-${agentId}`,
      labels: { app: 'yamtam-agent', tier, agentId },
    },
    spec: {
      restartPolicy: 'Never',
      containers: [{
        name:  'agent',
        image: `ghcr.io/yamtam/agent:latest`,
        env:   [
          { name: 'YAMTAM_AGENT_ID',   value: agentId },
          { name: 'YAMTAM_AGENT_TIER', value: tier },
        ],
        resources: {
          requests: { memory: '256Mi', cpu: '100m' },
          limits:   { memory: '512Mi', cpu: '500m' },
        },
        securityContext: {
          readOnlyRootFilesystem: true,
          allowPrivilegeEscalation: false,
          runAsNonRoot: true,
        },
      }],
    },
  })
  return pod.body.metadata!.name!
}

Scale a Deployment

async function scaleDeployment(name: string, namespace: string, replicas: number) {
  await appsApi.patchNamespacedDeploymentScale(name, namespace,
    [{ op: 'replace', path: '/spec/replicas', value: replicas }],
    undefined, undefined, undefined, undefined, undefined,
    { headers: { 'Content-Type': 'application/json-patch+json' } }
  )
  console.log(`[k8s] scaled ${name} → ${replicas} replicas`)
}

Stream pod logs

async function streamPodLogs(podName: string, namespace = 'yamtam') {
  const logStream = await new k8s.Log(kc).log(
    namespace,
    podName,
    undefined,    // container name (undefined = first)
    process.stdout,
    {
      follow:    true,
      tailLines: 100,
      timestamps: true,
    }
  )
  return logStream  // call logStream.destroy() to stop
}

Wait for Pod to be Running

async function waitForPod(podName: string, namespace = 'yamtam', timeoutMs = 60_000) {
  const start = Date.now()

  while (Date.now() - start < timeoutMs) {
    const pod   = await coreApi.readNamespacedPod(podName, namespace)
    const phase = pod.body.status?.phase

    if (phase === 'Running')  return true
    if (phase === 'Failed')   throw new Error(`[k8s] pod ${podName} failed`)
    if (phase === 'Succeeded') return true

    await new Promise(r => setTimeout(r, 2000))
  }
  throw new Error(`[k8s] pod ${podName} not ready after ${timeoutMs}ms`)
}

Anti-Fake-Pass Checklist

❌ loadFromDefault() outside cluster without KUBECONFIG set → uses ~/.kube/config, may hit wrong cluster
❌ No resource limits on spawned pods → runaway agent exhausts node memory
❌ readOnlyRootFilesystem without emptyDir volume → agent can't write temp files
❌ Watching pods without timeout → watch connection drops after ~1h; reconnect required
❌ JSON patch without Content-Type: application/json-patch+json → K8s rejects with 415
❌ Deleting pod without waiting → successor pod starts with stale ConfigMap if not flushed

04 同仓库更多 Skills

Horonomos-laps
Horonomos-laps/Yana-AI

openai--sentry--sentry

Use when the user asks to inspect Sentry issues or events, summarize recent production errors, or pull basic Sentry health data via the Sentry API; perform read-only queries with the bundled script and require `SENTRY_AUTH_TOKEN`.

★ 13
Horonomos-laps
Horonomos-laps/Yana-AI

ito-basket-compare

Compare Itô prediction-market baskets against a user's knowledge base, portfolio notes, financial context, watchlist, or research thesis. Use for read-only basket comparison and gap analysis without investment advice or live trading.

★ 13
Horonomos-laps
Horonomos-laps/Yana-AI

implementing-network-traffic-baselining

Build network traffic baselines from NetFlow/IPFIX data using Python pandas for statistical analysis, z-score anomaly detection, and hourly/daily traffic pattern profiling

★ 13
Horonomos-laps
Horonomos-laps/Yana-AI

detecting-evasion-techniques-in-endpoint-logs

Detects defense evasion techniques used by adversaries in endpoint logs including log tampering, timestomping, process injection, and security tool disabling. Use when investigating suspicious endpoint behavior, building detection rules for evasion tactics, or conducting threat hunting for stealthy adversary activity. Activates for requests involving evasion detection, defense evasion analysis, log tampering detection, or MITRE ATT&CK TA0005.

★ 13
Horonomos-laps
Horonomos-laps/Yana-AI

implementing-microsegmentation-with-guardicore

Implementing microsegmentation using Akamai Guardicore Segmentation to map application dependencies, create granular network policies, visualize east-west traffic flows, and enforce least-privilege communication between workloads across data centers and cloud.

★ 13
Horonomos-laps
Horonomos-laps/Yana-AI

ramdisk-tmpfs-patterns

暂无简介

★ 13