agentleFS
kyle-chalmerskyle-chalmers/streamsnow/skills/onboard/SKILL.md

onboard

Skill17 starsChanged 32 days agoMITView on GitHub

Get this machine, this repo and its Snowflake account ready for StreamSnow. Detects what is already done and does only what is missing, so it is safe to re-run to ask "am I set up?". Use when the user says "set me up", "onboard", "I just cloned this repo", "is my machine ready", "add StreamSnow to this repo", or "set up Snowflake for StreamSnow", and when /build-app finds the machine or repo isn't ready.

165 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use soak-test, team-live-ops and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • lilangMaxSkill

    soak-test

    Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.

    Also in 9 projects that use this

  • lilangMaxSkill

    team-live-ops

    Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.

    Also in 9 projects that use this

  • HK-hubSkill

    teach-impeccable

    One-time setup that gathers design context for your project and saves it to your AI config file. Run once to establish persistent design guidelines.

    Also in 19 projects that use this

  • CoralGameSkill

    project-stage-detect

    Automatically analyze project state, detect stage, identify gaps, and recommend next steps based on existing artifacts.

    Also in 15 projects that use this

  • CoralGameSkill

    team-level

    Orchestrate level design team: level-designer + narrative-director + world-builder + art-director + systems-designer + qa-tester for complete area/level creation.

    Also in 11 projects that use this

  • +5 more from lilangMax
  • +2 more from CoralGame

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 153 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before each action, one line of what and why.

    line 20
  2. Distinctive:

    Every question says what it decides and what each answer changes.

    line 22
  3. Distinctive:

    Add the plugin to the repo's Claude settings when missing (setup.md §2e).

    line 38
  4. Distinctive:

    no Snowflake connection: set it up for them, or guide them (setup.md §2a).

    line 49
  5. Very specific:

    No config, no apps: streamsnow configure with the confirmed flags, then streamsnow init --no-starter-app (setup.md §2c).

    line 59

“Common” is measured: kept by at least 6 in 10 of the 153 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/onboard

Repo overlay: if .streamsnow/overlays/onboard.md exists in this repo, read it first: committed, repo-specific additions and overrides (_shared/overlays.md). Outside Claude Code, also read _shared/other-agents.md.

Gets this machine, this repo and its Snowflake account ready to build and ship apps. Everything that needs no answers runs first, then one round of questions, then the build, then the steps that wait on other people. Re-running re-checks and does only what is missing. The trust rules at the top of setup.md apply throughout.

Narrate every step

  • Open each stage with one line: what it does, and whether it will ask anything.
  • Before each action, one line of what and why; after it, one line with the result. Actions that need no approval are narrated too.
  • Every question says what it decides and what each answer changes. Setup questions follow the question rule in Stage 2.
  • Close each stage with what got done and what is next. Plain words, no walls of text.

Stage 1 · Detect and prepare (no questions except one install approval)

  1. Native Windows outside WSL: stop and follow setup.md §0.
  2. Detect: streamsnow doctor --format json (no CLI: setup.md §0b first), whether streamsnow.config.yaml exists, whether apps/*/streamlit_app.py exists, whether .claude/settings.json enables streamsnow@streamsnow, and the Snowflake access inventory (setup.md §2a). Report the state in one line.
  3. With any Snowflake access: the read-only investigation (setup.md §2b), including the admin check (setup.md §2d step 1).
  4. One batched approval for every missing tool, then install one at a time and re-run doctor after each (setup.md §0b and §1). If snow was just installed or fixed, redo step 3.
  5. The project Python environment (setup.md §1c) and the browser check (setup.md §1b).
  6. Add the plugin to the repo's Claude settings when missing (setup.md §2e).

Nothing in Stage 1 writes streamsnow.config.yaml or the governed repo files.

Stage 2 · One round of questions

The question rule. Explain each setting immediately before its separate question. Use the host agent's question interface when it has one (Claude Code: AskUserQuestion, up to 4 per call); otherwise ask in chat and wait for an explicit answer. Detection never counts as confirmation: what Stage 1 found is the recommended first option. With a config already present nothing is being decided, so skip the setup questions (setup.md §2c). In order:

  • no Snowflake connection: set it up for them, or guide them (setup.md §2a); then redo Stage 1 step 3;
  • one question each: runtime, sources, app-data schema, denied schemas, deploy source, and the connection when §2a has not confirmed it (setup.md §2c);
  • the defaults it does not ask: bullets explaining each, then one keep or change question (§2c);
  • git name and email, only if missing;
  • apps but no config: confirm the adopt plan (adopt.md);
  • admin setup not confirmed: explain the script, then ask who runs it (setup.md §2d).

Stage 3 · Build (after the user confirms)

  • No config, no apps: streamsnow configure with the confirmed flags, then streamsnow init --no-starter-app (setup.md §2c).
  • Apps, no config: the adopt flow (adopt.md).
  • Config present and doctor's repo-files lists missing files: init --no-starter-app writes only those.
  • A non-default connection: choose how local preview reads it (setup.md §3).
  • Then pre-commit install, and doctor until every required row passes. Show the tree and the result.

Stage 4 · Finish what waits on others

  1. Snowflake admin setup when not confirmed (setup.md §2d): streamsnow ci-key create, the admin file in .internal/, the hand-off. Never run the admin SQL.
  2. Once the admin objects are confirmed: offer ci-key verify, then streamsnow ci-key push (setup.md §2e).
  3. With a GitHub remote, whether or not step 2 ran: read whether GitHub deletes merged branches and, if it does not, ask yes or no before turning it on (setup.md §2e).
  4. Offer to commit the new repo files and .claude/settings.json.

Done when

Name the end state:

  • Ready to build and preview: Stages 1 to 3 pass. Next: /build-app.
  • Ready to deploy: Stage 4 passes too.

With only the first, name what is pending ("waiting on your Snowflake admin") and say that re-running /onboard re-checks.

Out of scope

Building an app: /build-app. Porting one: /migrate-app. Deploying: /ship-app.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.