latent-sre/homelab-agents/onboarding-map
- community
Checking that the SKILL.md file is still in the repository.
Maps the fleet's explicit-only onboarding workflows without running them: which exist (sde-agents:service-onboard for a new or ad-hoc service, sde-agents:host-onboard for a new or rebuilt machine), the order — host first, then the services on it — and how to reach each on the current host. Use when someone asks whether an onboarding checklist or workflow exists, how to start onboarding a service or a host, or states new-service or new-host intent on a host where no routing agent is reachable. Recommending is not executing: the checklists run under sde-agents:homelab-engineer's change tiers, and this map changes nothing and grants nothing. Not for performing the onboarding or any lab change — use sde-agents:homelab-engineer — and not for auditing an existing setup, which is sde-agents:lab-audit for hygiene or sde-agents:security-audit for the adversary's view.
Install
npx vskill@latest install latent-sre/homelab-agents/onboarding-mapSKILL.md
A map, not a procedure: which onboarding workflows exist, which one this situation needs, and how to reach it here. It holds no authority and carries no steps — everything it names runs under sde-agents:homelab-engineer's change tiers, so naming a workflow never starts one and never supplies an approval.
Which workflow applies
| Situation | Route | Why | |
|---|---|---|---|
| A machine new to the lab, or rebuilt | sde-agents:host-onboard | Host baseline plus role-dependent protection before workloads depend on it; backups apply to irreplaceable state, and access changes require recovery proof | |
| A service being added, or an ad-hoc one brought up to standard | sde-agents:service-onboard | Placement, config as code, storage, network, security, monitoring, an operating doc, and an end-to-end verify | |
| A new machine that will also run a new service | Host first, then the service — two sequenced lifecycles | The service checklist assumes the machine under it is already a lab citizen; merging them loses that assumption and fans one request into two efforts at once | |
| A service landing on a host the lab already manages | The service workflow alone | An already-onboarded host does not repeat host onboarding — proportional service checks are the whole job | |
| "Is this set up correctly?", with nothing to change | sde-agents:lab-audit for hygiene, sde-agents:security-audit for the adversary's view | A read-only question is an audit; opening a checklist that applies live changes answers a question nobody asked | |
| Application code, or lab-shaping architecture | sde-agents:sde-fullstack; up the sde-agents:eng-ladder | Not onboarding at all — shared vocabulary ("add", "deploy", "service") is not shared remit |
Four states — say which one this is
| State | What it means | What it authorizes | |
|---|---|---|---|
| Discovery | the workflow exists, and this map names it | nothing | |
| Recommendation | it applies to this request, and here is why | nothing | |
| Activation | its checklist is open under sde-agents:homelab-engineer | that agent's Tier 0 observation and Tier 1 preparation | |
| Execution | a step reaches a live target | the live outcome the operator authorized within its scope, limits, and that agent's host controls |
Discovery and recommendation are free; activation and execution are not. Collapsing them fails in both directions — a user who never learns the workflow exists, and a workflow that reads as running because something named it. A pause during activation or execution belongs to a gate, and sde-agents:homelab-engineer names which one.
Reaching one from here
Plain language to sde-agents:homelab-engineer — "add this service to my lab", "onboard this new VM" — is the intended route: it owns change authority and works both checklists under its tiers. Naming the workflow directly, /sde-agents:host-onboard or /sde-agents:service-onboard, is the fallback, and it is the only route on a host that neither delegates to an agent by description nor lists explicit-only workflows to the model. The sigil differs by host, and the copy of this map installed on a host is written in that host's own form — so when the plain-language route does not surface these workflows, say the workflow's name exactly as it appears above.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/onboarding-map/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/latent-sre/homelab-agents/onboarding-map)