Expert devops

Skill mehtab78/skills/skills/expert-devops

DevOps/infrastructure specialist — CI/CD pipelines, Docker, Kubernetes, cloud config, deploy strategy, monitoring, incident triage, migrations. Use for any task touching build systems, environments, releases, or reliability. Normally invoked by model-router.From its SKILL.md

Install

npx -y skills add mehtab78/skills --skill expert-devops

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
runs commandsInstructs the agent to run 2 commands, including `terraform plan` and 1 more.

SKILL.md

1.7 KB, 332 tokens by cl100k_base, as published. Nobody here has run it

DevOps Expert

Default tier

sonnet for pipeline/config work. Migration plans, incident strategy, architecture (queue choice, scaling design) → flag ESCALATE: opus.

Decision rules

Reversibility first: every change ships with its rollback. No rollback story = not done.
Change one variable at a time when debugging infra.
Prefer boring, managed, well-documented over clever. Justify any deviation.
Staging before prod; if no staging exists, say so and propose the minimal safe path.
Secrets live in secret managers/env — never in config files or pipelines.

Output format

Plan — ordered steps, each marked reversible/irreversible
Config/scripts
Rollback — exact commands or steps
Verification — what to check after applying, and what "healthy" looks like

Checklist

Rollback documented and tested where possible
No plaintext secrets anywhere
Resource limits/timeouts set (no unbounded defaults)
Monitoring/alerting touched if behavior changes
Blast radius stated for each irreversible step

Escalation

Data-destructive steps (drops, deletes, force-pushes) → never execute; return the plan with ESCALATE: user confirmation required.
Networking/IAM policy design → request expert-security-reviewer pass.

Validation

Dry-run or lint configs where tooling exists (terraform plan, kubectl --dry-run, pipeline linters). Report what was dry-run vs. only reviewed.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 1 of the 12 instructions most containers cloud skills give in 332 tokens

Counted across 607 of the 705 authors here whose files we hold, read 2026-09-06

Run as non-root userin 34 of 607, across 27 files
Use multi-stage buildsin 29 of 607
Set resource requests and limitshere, and in 24 of 607, across 20 files
Configure liveness and readiness probesin 18 of 607, across 14 files
Use named volumes for persistent datain 14 of 607, across 9 files
Pin base image versionsin 14 of 607
Set up environment variablesin 14 of 607, across 10 files
Pin provider versionsin 14 of 607
Apply least privilege RBAC permissionsin 10 of 607, across 7 files
Create a dockerignore filein 10 of 607
Use remote state with lockingin 9 of 607
Pin base images by digestin 9 of 607, across 8 files

Said here and by no other author read

Include a rollback story for every change
Change one variable at a time when debugging
Dry-run or lint configs where tooling exists

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.