agentleFS
samwelmungasamwelmunga/jenga-npm/.agents/skills/j-cloud-connect/SKILL.md

j.cloud-connect

Skill0 starsChanged 22 days agoApache-2.0View on GitHub

Guided cloud storage setup wizard — installs rclone if missing, lets you pick any rclone-supported backend from a live provider list, runs that backend's own config/auth flow, and independently verifies the remote works.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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:

    If the user already stated a specific backend and remote name in their request (e.g.

    line 62
  2. Very specific:

    If the user only wants to see what backends are available without configuring one yet, run bash …

    line 68

≈ “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.

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

Cloud Connect — Guided rclone Setup Wizard

Purpose

E60_S01 builds a single guided flow for connecting a cloud storage backend via rclone: install (or detect) rclone, pick a backend from rclone's own live provider list, run that backend's own rclone config create prompts (surfacing any auth URL directly), wait for explicit user confirmation, then independently verify the remote works. All of that behavior already lives in two standalone scripts — scripts/install-rclone.sh (E60_S01_T01) and scripts/configure-backend.sh (E60_S01_T02) — each fully self-contained, argument-driven, and independently tested. Per this repo's "Scripts Over Inline Logic" principle (CLAUDE.md), this SKILL.md introduces no new install, selection, config, auth, or verification logic of its own — it is a thin entry point that sequences those two scripts and relays their output.

Instructions

  1. Run the install/detect step first. Invoke:

    bash skills/j-cloud-connect/scripts/install-rclone.sh

    Relay its stdout/stderr to the user as-is (it already produces clear, human-readable status — e.g. "rclone is already installed at ...", or an actionable error for an unsupported OS, a missing package manager, or a network failure). Do not reinterpret or summarize away its output.

    • Exit 0: rclone is confirmed on PATH and runnable (whether it was already present or was just installed). Proceed to step 2.
    • Non-zero exit: installation failed or could not be verified. Stop here — do not proceed to step 2 — and surface the script's error message to the user unchanged. It already names the specific cause (unsupported OS, missing Homebrew/curl/sudo, network failure, post-install verification failure) and, where applicable, a manual-install fallback URL.
  2. Run the backend selection/config/auth/verify step. Invoke:

    bash skills/j-cloud-connect/scripts/configure-backend.sh

    with stdio fully inherited/interactive — this script prompts the user directly (backend number, remote name, an explicit yes/no confirmation once any backend auth flow has run, e.g. opening an OAuth URL) and must be allowed to do so; do not pre-supply answers or short-circuit its prompts.

    • If the user already stated a specific backend and remote name in their request (e.g. "connect my Google Drive as gdrive"), you may pass them positionally instead of relying on the interactive prompts: bash skills/j-cloud-connect/scripts/configure-backend.sh <backend-type> <remote-name> — the script validates <backend-type> against the live provider list itself and fails clearly if it doesn't match. If the requested backend name doesn't resolve, fall back to running the script with no arguments so the user can pick from the live menu.
    • If the user only wants to see what backends are available without configuring one yet, run bash skills/j-cloud-connect/scripts/configure-backend.sh --list instead and report the printed menu — it is a read-only, no-side-effect mode.
    • Relay all of the script's output to the user as-is, including any auth URL rclone's own config create flow prints — never paraphrase, hide, or re-derive that URL.
    • Exit 0: the script itself already printed a PASS: remote '<name>' is configured and verified working. line plus rclone about output — relay that verbatim as the final result.
    • Non-zero exit: the script already printed the specific failure reason (invalid backend selection, rclone config create failure, a declined confirmation, or a FAIL: verification line from rclone about) — relay it unchanged rather than guessing at the cause.
  3. Do not run step 2 if step 1 failed. The wizard is a strict sequence: configure-backend.sh itself assumes rclone is already on PATH and will die immediately with a pointer back to install-rclone.sh if it isn't, so there is nothing to gain by invoking it after a failed install — surface the step 1 failure and stop.

Out of Scope

  • Any install, detection, or OS-specific package-manager logic — that is entirely scripts/install-rclone.sh's scope (E60_S01_T01). Do not duplicate or reimplement any part of it here.
  • Any backend menu sourcing, rclone config create invocation, auth-confirmation prompt, or post-confirmation rclone about verification — that is entirely scripts/configure-backend.sh's scope (E60_S01_T02). Do not duplicate, hardcode a backend list, or reimplement any part of it here.
  • Per-backend bespoke guidance (e.g. special-casing Google Drive's auth flow beyond what rclone's own config create already prints) — the story's acceptance criteria require the identical flow for every backend; any such special-casing belongs to neither script nor this SKILL.md.

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.