deployment-decisions
Skill0 starsChanged 4 months agoMITView on GitHub
Use when the user wants to deploy but the method, platform, or process is not yet established.
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- templates/AGENTS.mdAGENTS.md
- branch-safelySkill
- deploySkill
- deploy-vercelSkill
- deploy-vpsSkill
- documentationSkill
- e2e-testingSkill
- feature-testingSkill
- monitoringSkill
- orchestrateSkill
- quality-reviewSkill
- requirements-gatheringSkill
- solution-designSkill
- task-breakdownSkill
- ui-uxSkill
- unit-testingSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
16 rules. About 0 are common, 13 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈13
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Ask: Do you already have a way to deploy this project?
- Very specific: line 14
If yes: ask what it is (GitHub Actions, a script, a service, manual steps).
- Distinctive: line 15
If no: go to step 2.
- Distinctive: line 19
Ask: Do you have a server or VPS?
- Distinctive: line 21
If yes: go to VPS path (step 3a).
- Distinctive: line 22
If no: go to step 3b.
- Distinctive: line 26
Ask: Do you already have Docker set up on the server?
- Distinctive: line 28
If yes: suggest GitHub Actions + Docker.
- Distinctive: line 29
If no: suggest installing Docker first, then the same workflow.
- Distinctive: line 48
If the user says "I don't know" or "just pick one": choose the simplest free option for their stack and explain why in one sentence.
- Distinctive: line 52
After the user chooses:
- Very specific: line 62
If the user wants, also save to ~/.agents/skills/deploy-project/SKILL.md for global reuse.
- Distinctive: line 66
Never suggest paid services unless the user explicitly asks.
- Very specific: line 67
Never handle tokens, passwords, or API keys directly.
- Distinctive: line 68
Keep the setup under 5 steps.
- Distinctive: line 69
If something fails, diagnose the specific step before suggesting fixes.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Deployment Decisions
Guide the user through choosing how to deploy. Keep it simple. Ask one question at a time. Never assume knowledge. Prefer free options. Save the chosen process as a project skill for future use.
Step 1 — Check for an existing process
Ask: Do you already have a way to deploy this project?
- If yes: ask what it is (GitHub Actions, a script, a service, manual steps). Verify it works by asking how they normally run it. Save it as a project skill at
.agents/skills/deploy-project/SKILL.mdso future runs reuse it. Done. - If no: go to step 2.
Step 2 — Assess infrastructure
Ask: Do you have a server or VPS? (for example a DigitalOcean droplet, Hetzner, AWS EC2, a home server)
- If yes: go to VPS path (step 3a).
- If no: go to step 3b.
Step 3a — VPS path
Ask: Do you already have Docker set up on the server?
- If yes: suggest GitHub Actions + Docker. Create a GitHub Actions workflow that builds a Docker image and deploys to the VPS via SSH. This is simple, free (GitHub Actions has generous free minutes), and repeatable. Save as project skill.
- If no: suggest installing Docker first, then the same workflow. Walk them through Docker install on their server (one command per distro).
Credentials: use GitHub Actions secrets. The agent generates the workflow file; the user adds secrets (VPS_HOST, VPS_USER, VPS_SSH_KEY) in the GitHub repo settings. The agent never handles these directly.
Step 3b — No server
Present the free options with one-line pros and cons:
| Platform | Pros | Cons |
| Vercel | Zero config for Next.js/React, free tier, automatic HTTPS, instant deploys | Free tier limits, framework-focused |
| Netlify | Similar to Vercel, generous free tier, form handling | Same framework bias |
| Cloudflare Pages | Fast, free, edge network | Less flexible for backends |
| GitHub Pages | Free, simple | Static only, no server-side |
| Railway | Free tier, supports any stack, simple | Free tier limited hours |
| Render | Free tier, Docker support | Cold starts on free tier |
Suggest the simplest option that fits their stack. For a Next.js app, suggest Vercel. For a static site, suggest Cloudflare Pages or GitHub Pages. For a full-stack app with a database, suggest Railway or Render.
If the user says "I don't know" or "just pick one": choose the simplest free option for their stack and explain why in one sentence.
Step 4 — Set up and save
After the user chooses:
- Walk through the setup step by step, one command or action at a time.
- For services (Vercel, Netlify, etc.): guide them to create an account and connect their repo. Prefer
ghCLI or the service's own CLI for auth — the agent should not handle tokens directly. - For VPS: generate the GitHub Actions workflow and any Dockerfiles needed.
- Save the deployment process as
.agents/skills/deploy-project/SKILL.mdwith:- Platform chosen
- How to deploy (command or push-to-deploy)
- Any required environment variables (names only, never values)
- How to verify deployment worked (URL or health check)
- If the user wants, also save to
~/.agents/skills/deploy-project/SKILL.mdfor global reuse.
Rules
- Never suggest paid services unless the user explicitly asks.
- Never handle tokens, passwords, or API keys directly. Use
ghCLI, service CLIs, or tell the user what to add in their dashboard. - Keep the setup under 5 steps. If it needs more, simplify.
- If something fails, diagnose the specific step before suggesting fixes.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

