k8s

Operate the joelclaw Kubernetes cluster — Talos Linux on Colima (Mac Mini). Deploy services, check health, debug pods, recover from restarts, add ports, manage Helm releases, inspect logs, fix networking. Triggers on: 'kubectl', 'pods', 'deploy to k8s', 'cluster health', 'restart pod', 'helm install', 'talosctl', 'colima', 'nodeport', 'flannel', 'port mapping', 'k8s down', 'cluster not working', 'add a port', 'PVC', 'storage', any k8s/Talos/Colima infrastructure task. Also triggers on service-specific deploy: 'deploy redis', 'redeploy inngest', 'livekit helm', 'pds not responding'.

Quality

85%

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

71%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.

This is a highly actionable operations reference with exceptional concrete detail — real commands, manifests, ports, and validation-minded recovery rules. Its weaknesses are length and structure: dated incident narratives pad the main body, and large topics (agent runner, NFS, Danger Zones) should be split into the existing one-level reference pattern rather than inlined in SKILL.md.

Suggestions

Move dated incident narratives (2026-03-17, 2026-04-15, 2026-05-30) and the longest Danger Zone entries (#16, #19, #20) into a references/incidents.md or 'old patterns' section, keeping only the durable rule and one-line pointer in SKILL.md.

Split the Agent Runner contract and the NAS NFS section into their own reference files (e.g. references/agent-runner.md, references/nas-nfs.md), leaving a short summary plus the env-var table pointer in the body.

Condense the 20 Danger Zones to one-line rules in a table with a pointer to details, mirroring how operations.md already handles port mappings and recovery.

Dimension	Reasoning	Score
Conciseness	The body is dense with genuinely non-obvious operational knowledge (no explaining concepts Claude already knows), but the incident narratives anchored to specific dates ('2026-03-17 incident', '2026-04-15 incident', '2026-05-30 reboot') and long narrative Danger Zone entries (#16, #19, #20) are padded and could be tightened. Not a 2 because most content earns its place; not a 4 because the verbosity is more than minor — dated time-sensitive detail is embedded in the main body rather than a deprecated/old-patterns section.	3 / 5
Actionability	Fully executable throughout: copy-paste-ready commands (Redis AOF fix pod manifest, talosctl/kubectl config commands, deploy scripts), exact ports, paths, and curl health endpoints covering the common operational cases. Anchor 4 ('minor gaps') doesn't apply — specific examples cover the common cases.	5 / 5
Workflow Clarity	Recovery sequences with validation gates are present: the Colima crash-loop recovery steps, the Redis AOF fix with an explicit logs-check step, Danger Zone #20's durable recovery sequence ending in 'Verify localhost:3838, 8108, ... are real Docker/Lima listeners', and the ADR-0244 durable-recovery rule defining a post-restart stability window. Not a 5 because procedures are scattered across sections rather than forming coherent sequences, and some destructive paths (e.g. PVC purge) defer their validation detail to prose rules; not a 3 because validation for destructive/batch operations is emphatically present.	4 / 5
Progressive Disclosure	The bundle is well-formed: references/operations.md is real, one level deep (no nested references inside it), and clearly signaled three times from the body. However, the ~520-line body inlines large blocks that clearly belong in separate reference files — the full agent-runner runtime contract, the NFS deep-dive, and 20 Danger Zones — leaving SKILL.md far heavier than an overview. Not a 2 because section structure is good and the reference is not buried; not a 4 because the inlined bulk goes beyond 'minor organization gaps'.	3 / 5
	Total	15 / 20 Passed

Description

96%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.

The description is a strong example: third-person voice, concrete action list for both 'what' and an explicit 'Triggers on' clause for 'when', and rich natural-language trigger phrases including service-specific ones. The only weakness is a handful of broad single-word triggers ('pods', 'storage', 'PVC') that could overlap with other infrastructure skills.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Deploy services, check health, debug pods, recover from restarts, add ports, manage Helm releases, inspect logs, fix networking' — with comprehensive coverage of the cluster-operations domain. Anchor 4 ('minor gaps in coverage') doesn't apply since nothing listed is generic or missing.	5 / 5
Completeness	Explicitly answers both questions: the 'what' via the concrete action list and the 'when' via 'Triggers on: ... any k8s/Talos/Colima infrastructure task. Also triggers on service-specific deploy'. Not a 4, since the 'when' clause is fully explicit with concrete trigger phrases.	5 / 5
Trigger Term Quality	Comprehensive natural trigger phrases users would actually say: 'k8s down', 'cluster not working', 'add a port', 'restart pod', 'deploy redis', 'pds not responding', plus tool names (kubectl, talosctl, colima, helm). Anchor 4 requires 'a few natural terms missing', which isn't the case here.	5 / 5
Distinctiveness Conflict Risk	Clearly niched to the joelclaw cluster (Talos Linux on Colima, Mac Mini) with mostly distinct triggers. Not a 5 because generic standalone triggers like 'pods', 'storage', and 'PVC' carry minor overlap risk with any other Kubernetes-related skill; not a 3 since the domain scoping is unmistakable.	4 / 5
	Total	19 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (536 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: joelhooks/joelclaw
Path: skills/k8s/SKILL.md
Commit: aa0116c

Reviewed: 8 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.