Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/gke-labs/kube-agentsnpx agentmods add skills/gke-labs/kube-agents/gke-stockout-investigatorWrote 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/gke-labs/kube-agents/gke-stockout-investigator)<a href="https://agentmods.dev/skills/gke-labs/kube-agents/gke-stockout-investigator"><img src="https://agentmods.dev/badge/skills/gke-labs/kube-agents/gke-stockout-investigator/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/gke-labs/kube-agents/gke-stockout-investigator"><img src="https://agentmods.dev/badge/skills/gke-labs/kube-agents/gke-stockout-investigator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 287 Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.Fix: Restrict tool access to only the tools required for the skill's stated purpose. Use an explicit allowlist rather than granting blanket access.
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 | $0.00034 | $0.05394 |
| Opus 5 | $0.00017 | $0.02697 |
| Sonnet 5 | $0.00007 | $0.01079 |
| Haiku 4.5 | $0.00003 | $0.00539 |
Grade B, and why
gke-stockout-investigator scanned grade B with 2 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 today.
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.
Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
When running in a background/PubSub context or when a new SRE review Pull Request with remediation is being created, before providing your final response, you MUST call the `send_notification` tool to notify the user/SRE Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
> In background/PubSub sessions, any invocation of `execute_code` (Python or bash script execution) triggers interactive command approval safeguards that will block and hang the session indefinitely. You MUST execute com How it starts
The opening of the file, as written. The whole thing — 304 lines — stays where its author put it; the contents beside it link to each section on GitHub.
GKE Stockout Diagnosis & GitOps Remediation Skill
This skill guides the Platform Agent on how to handle, diagnose, and remediate GKE capacity stockout alerts (scale.up.error.out.of.resources or equivalent) received via Google Cloud Pub/Sub.
Workflow
1. Notify User of Investigation Start (MANDATORY FIRST STEP)
Before executing ANY terminal commands, scripts, or diagnostics, you MUST immediately call the send_notification tool (or mcp_platform_control_send_notification) to announce that the stockout alert was received and investigation is starting:
send_notification(message="🚨 GKE Stockout Investigation Started\nWorkload: <workload_name>\nCluster: <cluster_name>\nDetails: A GKE capacity stockout alert is confirmed. I am starting an investigation and diagnosis.")
CRITICAL: You MUST invoke send_notification (or mcp_platform_control_send_notification) as an actual tool call. Do NOT output this notification as plain markdown text. After calling the tool, proceed immediately to Step 2 in your next tool turn without stopping.
[!IMPORTANT] When running in a background or PubSub context, NEVER use the
execute_codetool or write Python scripts/subshells, as they trigger command approval safeguards and block indefinitely waiting for human approval. Only execute standard command-line tools directly (kubectl,gh,gcloud) or dedicated tools likesend_notification.
2. Pre-Diagnosis Verifications (Duplicate PRs & False Signal Checks)
After sending the initial notification, perform two critical safety checks to see if you should stop immediately:
A. Check for Existing Relevant Pull Requests (Duplicate Prevention)
To prevent duplicate effort and redundant PRs, inspect currently open Pull Requests in the repository:
- Resolve the GitOps repository as
<owner>/<repo>, and name it on everyghcall from here on. The pod is not a git checkout (Step 3 explains why), soghhas nooriginremote to infer a repository from: an unqualifiedgh pr listfails on repository resolution rather than returning a list, and the duplicate check silently never happens. Query the$GITOPS_STATE_CONFIGMAPConfigMap:
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.
- today Changed · +10 lines 7f2b44458f57
- 9d ago Changed · +1 lines 81ca6031a604
- 12d ago First seen · 293 lines · 34 tokens per session scan B 7babea5cfbfa
gke-stockout-investigator is a skill published in the GitHub repository gke-labs/kube-agents (54 stars, last pushed today), licensed Apache-2.0. It adds 34 tokens to every session and 5,394 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it B with 2 findings (unrestricted tool access, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
gke-ai-troubleshooting-jobset-interruption
Diagnoses GKE JobSet interruptions, restarts, and preemptions for AI/ML training workloads autonomously. Use when troubleshooting JobSet restart loops, spot VM preemptions, node readiness failures, host VM issues, or coordinator worker crashes. Don't use for general GKE cluster creation, basic workload deployment, or…
gke-workload-troubleshooting
Diagnoses GKE workload failures (CrashLoopBackOff, OOMKilled, ImagePullBackOff, Pending, etc.) via logs and events. Use when pods fail to start or crash repeatedly. Don't use for GKE cluster infrastructure provisioning, node pool creation, or non-Kubernetes Google Cloud services.
gke-node-notready
Diagnoses GKE nodes reporting NotReady or Unknown status by inspecting node conditions, events, kubelet/containerd logs, and node metrics, then proposing safe remediations. Use when nodes show NotReady, when the kubelet stops posting node status, or when workloads are evicted or stuck Pending due to node health. Don't…
gke-ai-troubleshooting-handle-disruption-gpu-tpu
Diagnoses, predicts, and mitigates node disruptions during Compute Engine host maintenance and hardware or software maintenance events for GPU and TPU workloads on GKE. Use when diagnosing node disruptions, predicting host maintenance events on GPU/TPU nodepools, inspecting node interruption PromQL metrics, auditing…
agentcore-investigation
Investigate Bedrock AgentCore runtime sessions via CloudWatch Logs Insights — resolve session/trace IDs, query OTEL spans, filter noise, build timelines. Use when debugging AgentCore agent sessions, tracing tool calls, or analyzing latency.
troubleshoot-sandbox
Troubleshoot OpenSandbox issues by running diagnostics (logs, inspect, events, summary) via CLI or HTTP API to diagnose sandbox failures like OOM, crash, image pull errors, network problems, etc.