Conversation
The vast majority of users won't fork this and read the code — they'll
paste the GitHub URL into Claude.ai (web app) or open the cloned folder
with Claude Code, then ask in plain English for what they need. This
commit makes that path actually delightful.
Five project-level slash commands in .claude/commands/ — auto-discovered
by Claude Code. Each is a self-contained playbook Claude follows on the
user's behalf:
- /setup — first-time install walkthrough. 6 steps (verify Chrome, load
unpacked, get Gemini key, save key, optional rewrite rules, smoke test).
Asks for confirmation between each. Designed for someone who has never
visited chrome://extensions before.
- /configure-store <storefront-url> — visits the user's live storefront
with a browser tool, samples 3-5 products, observes the title pattern /
description structure / dialect / forbidden patterns, and emits three
paste-ready blocks (title rules, description rules, examples). The
output format is exactly what the Options page expects.
- /test-on <product-url> — 6-step copy + paste smoke test through the
extension. Handy for fresh installs and after rewrite-rules changes.
- /troubleshoot <symptom> — interactive triage. Verify version is
current → match symptom against the known-issues table from CLAUDE.md
→ request SW console diagnostics if no match. Side-effect-free.
- /release <patch|minor|major> — version bump in three places, CHANGELOG
entry, npm run check, commit, push, PR, wait for CI green, merge.
Marked disable-model-invocation:true because it has side effects;
maintainers invoke it explicitly.
CLAUDE.md gets a new "End-user playbook" section above the GitHub-rules
section. It maps casual user phrasings ("set me up", "configure for my
store at <url>", "fix this") to the right slash command (Claude Code) or
to the procedure body in the slash-command file (Claude.ai web app —
which doesn't run slash commands). Also lists 5 critical phrasing rules
for talking to non-technical users (no jargon, one question at a time,
confirm before destructive actions, show what success looks like, tell
them what NOT to share).
README front-loaded with a "🤖 Don't read all this — let Claude do it
for you" section right under the title. Contains:
- Option A (Claude.ai web app): paste the GitHub URL + a 4-row table
of canned prompts mapping common user goals to plain-English chat
messages.
- Option B (Claude Code CLI): the same 5 slash commands listed.
- Explicit "the rest of this README is for people who want to read the
code themselves" — non-tech users can stop right there.
.gitignore updated so .claude/* is ignored EXCEPT .claude/commands/,
.claude/agents/, and .claude/settings.json — the shared content that
should travel with the repo. The per-machine worktree state under
.claude/worktrees/ stays excluded.
All four CI checks (syntax + lint + version-stamps + manifest validity)
still green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This is the change that turns the repo into something a non-technical Salla store owner can actually use without reading any code. They paste the GitHub URL into Claude.ai or open the folder in Claude Code, ask in plain English, and Claude follows a documented playbook.
Slash commands (
.claude/commands/)Five project-level commands, auto-discovered by Claude Code in any fork:
/setupchrome://extensionsto a successful smoke test/configure-store <url>/test-on <url>/troubleshoot <symptom>/release <patch|minor|major>disable-model-invocation: true— maintainer-invoked only)CLAUDE.md additions
New "End-user playbook" section that maps casual user phrasings ("set me up", "configure for my store", "something's wrong") to the right slash command or, for Claude.ai web app users who don't have slash commands, the procedure body in the matching
.claude/commands/*.mdfile. Also adds 5 critical phrasing rules: no jargon, one question at a time, confirm before destructive actions, show success state, tell them what NOT to share (their API key).README front-loaded
Right under the title: "🤖 Don't read all this — let Claude do it for you". Two options:
https://github.com/OutQio/pilot+ a table of canned prompts they can copyExplicit "the rest of this README is for people who want to read the code" so non-tech users know they can stop reading.
.gitignorescopedWas
.claude/(excluding everything). Now.claude/*with!.claude/commands/,!.claude/agents/,!.claude/settings.jsonwhitelisted so the slash commands ship in the repo while the per-machine worktree state stays ignored.Test plan
npm run checkpasses locally (syntax + lint + version stamps)/setupand verify it auto-discovers🤖 Generated with Claude Code