chore: open-source setup — CI/CD, SDD, tests, release automation

.agents/skills/implement-story/SKILL.md

@@ -0,0 +1,14 @@
+---
+name: implement-story
+description: Implement an approved project story end to end and close its single Markdown record.
+argument-hint: Story ID, for example 009
+---
+
+# Implement Story
+
+1. Read `AGENTS.md`, `docs/stories/INDEX.md`, and the requested story. Continue only for `OPEN` or `IN_PROGRESS`.
+2. Set the story and index to `IN_PROGRESS`.
+3. Implement the smallest change that satisfies every acceptance criterion. Keep business logic in services or models and preserve existing changes.
+4. After each edit slice, run the cheapest focused format, build, or test check available.
+5. Update the story with completed criteria, outcome, tests, and validation. Set it and the index to `CLOSED` only when all criteria are met.
+6. Do not commit or push unless explicitly requested; use the story's Conventional Commit type if a commit is requested.

.agents/skills/new-story/SKILL.md

@@ -0,0 +1,14 @@
+---
+name: new-story
+description: Create and register a concise project story for a feature, fix, refactor, test, documentation, or configuration change.
+argument-hint: Describe the change
+---
+
+# New Story
+
+1. Read `AGENTS.md`, `docs/stories/INDEX.md`, and the relevant existing stories.
+2. Confirm the request is not already covered by an `OPEN` or `IN_PROGRESS` story.
+3. Choose the next numeric ID and a Conventional Commit type: `feat`, `fix`, `refactor`, `perf`, `docs`, `test`, `chore`, or `ci`.
+4. Copy `docs/stories/TEMPLATE.md` to `docs/stories/NNN-short-name.md` and fill it with status `OPEN`, intent, concise acceptance criteria, and validation expectations.
+5. Add it to `docs/stories/INDEX.md` in descending numeric order.
+6. Do not implement the story until it is approved.

.agents/skills/review-codebase/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: review-codebase
+description: Perform a read-only codebase review for bugs, risks, spec drift, test gaps, and candidate project stories.
+argument-hint: Optional review focus
+---
+
+# Review Codebase
+
+1. Inspect the current worktree and branch without modifying files or performing Git operations.
+2. Read `AGENTS.md`, `docs/stories/INDEX.md`, and relevant stories.
+3. Look for reproducible bugs, security or data risks, contract drift, CI gaps, and meaningful missing tests.
+4. Report findings first, ordered by severity, with file and line links. Distinguish evidence from assumptions.
+5. Suggest new stories only for work not already covered by an `OPEN` or `IN_PROGRESS` story. Do not create or implement them without approval.

.cursorrules

@@ -0,0 +1,16 @@
+# Cursor rules
+
+Agent guidance for this project lives in a single, provider-neutral source of truth.
+
+Read AGENTS.md at the repository root before making changes. It covers architecture,
+build/test/format commands, code style, testing policy, Conventional-Commit + model-attribution
+rules, and the Spec-Driven Development workflow.
+
+Highlights:
+- Conventional Commits are mandatory; PR titles must be valid Conventional Commits.
+- Business logic belongs in Services/ and Models/, never in SwiftUI views.
+- Use the Swift Testing framework; keep tests minimal and focused on real business logic.
+- Agent commits must include a `Generated-by: <model-id>` trailer (e.g. Generated-by: claude-opus-4-8).
+  Do NOT add Co-Authored-By: — the human user is the author.
+- Never include personal contact information (emails, phone numbers) in any file.
+- Do not add third-party dependencies; the app is intentionally dependency-free.

.editorconfig

@@ -0,0 +1,15 @@
+root = true
+
+[*]
+charset = utf-8
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true
+indent_style = space
+indent_size = 4
+
+[*.{yml,yaml,json,rb}]
+indent_size = 2
+
+[*.md]
+trim_trailing_whitespace = false

.githooks/commit-msg

@@ -0,0 +1,76 @@
+#!/bin/sh
+#
+# Mounty commit-msg hook:
+#   1. Enforce Conventional Commits.
+#   2. When Generated-by: is present, validate its format.
+#   3. Reject Co-Authored-By: lines that reference AI tools — use Generated-by: instead.
+# Install with: ./scripts/install-hooks.sh
+#
+set -eu
+
+msg_file="$1"
+
+# First non-comment, non-empty line = the subject.
+subject=$(grep -vE '^\s*#' "$msg_file" | grep -vE '^\s*$' | head -n1 || true)
+
+# Allow merge/revert/fixup commits through untouched.
+case "$subject" in
+    "Merge "*|"Revert "*|"fixup!"*|"squash!"*) exit 0 ;;
+esac
+
+# ── 1. Conventional Commits ───────────────────────────────────────────────────
+conventional='^(feat|fix|docs|style|refactor|perf|test|build|ci|chore|revert)(\([a-z0-9._/-]+\))?!?: .+'
+
+if ! printf '%s' "$subject" | grep -Eq "$conventional"; then
+    cat >&2 <<'EOF'
+✖ commit-msg: subject is not a valid Conventional Commit.
+
+  Expected: <type>[(scope)][!]: <description>
+  Types:    feat, fix, docs, style, refactor, perf, test, build, ci, chore, revert
+
+  Examples:
+    feat: add reconnect backoff
+    fix(mount): treat EEXIST as success
+    feat!: drop macOS 25 support
+
+See https://www.conventionalcommits.org/
+EOF
+    exit 1
+fi
+
+body=$(cat "$msg_file")
+
+# ── 2. Validate Generated-by: format when present ────────────────────────────
+if printf '%s' "$body" | grep -qiE '^Generated-by:'; then
+    # Must be exactly: Generated-by: <non-empty-model-id>
+    if ! printf '%s' "$body" | grep -qE '^Generated-by: [a-zA-Z0-9._+-]+$'; then
+        cat >&2 <<'EOF'
+✖ commit-msg: malformed Generated-by: trailer.
+
+  Expected:  Generated-by: <model-id>
+  Examples:  Generated-by: claude-sonnet-4-6
+             Generated-by: gpt-5
+             Generated-by: gemini-2.5-pro
+
+  The model id must be a single token on the same line, no extra text.
+EOF
+        exit 1
+    fi
+fi
+
+# ── 3. Reject Co-Authored-By: from AI tools ──────────────────────────────────
+if printf '%s' "$body" | grep -qiE '^Co-Authored-By:.*\b(Claude|GPT|Copilot|Gemini|Codex|Cursor|OpenAI|Anthropic)\b'; then
+    cat >&2 <<'EOF'
+✖ commit-msg: AI Co-Authored-By: trailer found. Use Generated-by: instead.
+
+  The human user is the commit author. Record the model as a tool, not a co-author:
+
+    Generated-by: claude-sonnet-4-6
+
+  Remove the Co-Authored-By: line and add Generated-by: instead.
+  See CONTRIBUTING.md for details.
+EOF
+    exit 1
+fi
+
+exit 0

.githooks/pre-commit

@@ -0,0 +1,29 @@
+#!/bin/sh
+#
+# Mounty pre-commit hook: auto-format staged Swift files with swift-format.
+# Install with: ./scripts/install-hooks.sh
+#
+set -eu
+
+# Resolve swift-format (prefer the Xcode toolchain version).
+if command -v swift-format >/dev/null 2>&1; then
+    FORMAT="swift-format"
+elif xcrun --find swift-format >/dev/null 2>&1; then
+    FORMAT="xcrun swift-format"
+else
+    echo "pre-commit: swift-format not found; skipping formatting." >&2
+    exit 0
+fi
+
+# Collect staged Swift files (added/copied/modified).
+staged_swift=$(git diff --cached --name-only --diff-filter=ACM -- '*.swift')
+[ -z "$staged_swift" ] && exit 0
+
+echo "pre-commit: formatting staged Swift files with swift-format…"
+printf '%s\n' "$staged_swift" | while IFS= read -r file; do
+    [ -f "$file" ] || continue
+    $FORMAT format -i "$file"
+    git add "$file"
+done
+
+exit 0

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,34 @@
+name: Bug report
+description: Report a problem with Mounty
+labels: ["bug"]
+body:
+  - type: textarea
+    id: what-happened
+    attributes:
+      label: What happened?
+      description: A clear description of the bug and what you expected instead.
+    validations:
+      required: true
+  - type: textarea
+    id: repro
+    attributes:
+      label: Steps to reproduce
+      placeholder: |
+        1. Add a volume smb://...
+        2. Enable automount
+        3. Toggle VPN
+    validations:
+      required: true
+  - type: input
+    id: versions
+    attributes:
+      label: Mounty & macOS version
+      placeholder: "Mounty 1.1.0, macOS 26.1"
+    validations:
+      required: true
+  - type: textarea
+    id: logs
+    attributes:
+      label: Relevant logs
+      description: "Console.app logs filtered by subsystem `ch.maptic.Mounty` (or `Mounty`)."
+      render: shell

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,21 @@
+name: Feature request
+description: Suggest an idea for Mounty
+labels: ["enhancement"]
+body:
+  - type: textarea
+    id: problem
+    attributes:
+      label: Problem
+      description: What problem would this feature solve? What are you trying to do?
+    validations:
+      required: true
+  - type: textarea
+    id: proposal
+    attributes:
+      label: Proposed solution
+    validations:
+      required: true
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives considered

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,22 @@
+<!--
+The PR title MUST be a valid Conventional Commit (e.g. "feat: add reconnect backoff").
+It becomes the squash-merge commit and drives the automated release.
+-->
+
+## What & why
+
+<!-- Describe the change and the motivation. Link the story in docs/stories/ if applicable. -->
+
+## Type of change
+
+- [ ] `fix` — bug fix (patch release)
+- [ ] `feat` — new feature (minor release)
+- [ ] breaking change (`!` / `BREAKING CHANGE:` — major release)
+- [ ] `docs` / `chore` / `refactor` / `test` / `ci` (no release)
+
+## Checklist
+
+- [ ] PR title is a valid Conventional Commit
+- [ ] Code is formatted (`swift-format`) and CI is green
+- [ ] Business logic changes are covered by tests
+- [ ] If AI-assisted, commits include a `Generated-by: <model-id>` trailer

.github/copilot-instructions.md

@@ -0,0 +1,17 @@
+# GitHub Copilot instructions
+
+This project's agent guidance is maintained in a single, provider-neutral source of truth.
+
+👉 **Read [`AGENTS.md`](../AGENTS.md)** at the repository root. It covers the architecture,
+build/test/format commands, code style, testing policy, Conventional-Commit + model-attribution
+rules, and the Spec-Driven Development workflow.
+
+Key reminders for generated code and commits:
+
+- Follow **Conventional Commits**; PR titles must be valid Conventional Commits.
+- Business logic goes in `Services/`/`Models/`, not in SwiftUI views.
+- Use the **Swift Testing** framework for tests; keep them minimal and meaningful.
+- Agent commits must include a `Generated-by: <model-id>` trailer (e.g. `Generated-by: gpt-5`).
+  Do **not** add `Co-Authored-By:` — the human user is the author.
+- Never include personal contact information (emails, phone numbers) in any file. Use only the
+  GitHub advisory URL for security reporting.

.github/workflows/ci.yml

@@ -0,0 +1,59 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+concurrency:
+  group: ci-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+jobs:
+  # Ensure the PR title is a valid Conventional Commit (it becomes the squash commit
+  # that drives release-please).
+  pr-title:
+    if: github.event_name == 'pull_request'
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: read
+    steps:
+      - uses: amannn/action-semantic-pull-request@v5
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+  format:
+    runs-on: macos-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Select Xcode
+        uses: maxim-lobanov/setup-xcode@v1
+        with:
+          xcode-version: latest-stable
+      - name: swift-format lint
+        run: xcrun swift-format lint --strict --recursive Mounty MountyTests
+
+  build-test:
+    runs-on: macos-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Select Xcode
+        uses: maxim-lobanov/setup-xcode@v1
+        with:
+          xcode-version: latest-stable
+      - name: Show toolchain
+        run: xcodebuild -version && swift --version
+      - name: Build & test
+        run: |
+          xcodebuild \
+            -project Mounty.xcodeproj \
+            -scheme Mounty \
+            -configuration Debug \
+            -destination 'platform=macOS' \
+            CODE_SIGNING_ALLOWED=NO \
+            SWIFT_TREAT_WARNINGS_AS_ERRORS=YES \
+            GCC_TREAT_WARNINGS_AS_ERRORS=YES \
+            clean test