hotsheet 0.17.0-beta.16 → 0.17.0-beta.19

package/README.md

@@ -10,9 +10,9 @@
 
 <br>
 
-**Hot Sheet** is a lightweight project management tool that runs entirely on your machine. Create tickets with a bullet-list interface, drag them into priority order, and your AI tools automatically get a structured worklist they can act on.
+**Hot Sheet** is a local-first project management tool wired into your AI coding workflow. Create tickets with a bullet-list interface, drag them into priority order, and your AI tools automatically get a structured worklist they can act on. A real PTY-backed terminal lives in the footer drawer — switch to the dashboard view to see every terminal across every project as a tile grid — so you can keep an eye on dev servers, tests, and the AI's own Claude Code session from the same window.
 
-No cloud. No logins. No JIRA. Just tickets and a tight feedback loop.
+No cloud. No logins. No JIRA. Just tickets, terminals, and a tight feedback loop.
 
 <br>
 
@@ -96,20 +96,36 @@ The loop stays tight because the AI always knows what to work on next — and yo
   <img src="docs/demo-5.png" alt="Multiple tickets selected with the batch toolbar and context menu" width="900">
 </p>
 
-**Detail panel** — side or bottom orientation (toggle in the toolbar), resizable, collapsible. Shows category, priority, status, and Up Next in a compact grid, plus title, details, tags, attachments, and editable notes. Click a note to edit inline; right-click to delete.
+**Detail panel** — side or bottom orientation (toggle in the toolbar), resizable, collapsible. Shows category, priority, status, and Up Next in a compact grid, plus title, details, tags, attachments, and editable notes. Click a note to edit inline; right-click to delete. **Reader mode** (book icon) opens any note — or the details body — in a distraction-free overlay that you can step through with chevrons or arrow keys without re-clicking the icon.
 
 <p align="center">
   <img src="docs/demo-6.png" alt="Detail panel in bottom orientation showing ticket details, tags, and notes" width="900">
 </p>
 
+**Embedded terminal** — a real PTY-backed terminal lives in the footer drawer next to the commands log. Configure multiple named terminals per project (dev server, tests, build, a Claude session, anything else), each with its own theme, font, and per-terminal shell-history pool — up-arrow on the dev-server tab won't surface the commit messages you typed on the deploy tab. **11 curated themes** (Default, Dracula, Solarized Dark/Light, Nord, Gruvbox, Monokai, One Dark, Tomorrow Night, GitHub Dark/Light) and 11 monospaced font choices, settable as a project default OR per-terminal via the gear popover. Right-click tabs to close, drag-to-reorder, single + button to spawn a one-off dynamic terminal. A macOS-Terminal.app-style quit-confirm catches Cmd+Q when long-running processes (claude, npm dev, anything heavier than an idle shell) are alive so you don't lose them by accident.
+
+**Shell integration that earns its keep** — xterm.js renders the buffer and Hot Sheet parses every standard shell-integration escape: **OSC 7** drives a cwd-aware toolbar chip you can click to open the working directory in Finder/Explorer; **OSC 8** makes hyperlinks emitted by `gh`, `cargo`, `ls --hyperlink=auto`, etc. actually clickable; **OSC 9** fires desktop toasts AND native OS notifications (system Notification Center / tray) when the window is backgrounded; **OSC 133** marks every prompt with a gutter glyph, supports `Cmd+Up`/`Cmd+Down` to jump between commands, and a hover popover with **Copy command** / **Copy output** / **Rerun** / **Ask Claude about this** buttons. A bell character (`\x07`) bounces the terminal tab and persists an accent outline until you visit it — and bells fired in **another project's** terminal surface as a dot on that project's tab so you don't miss them.
+
+**Terminal dashboard** — click the `square-terminal` button in the toolbar to enter a dedicated full-window view showing every terminal across every registered project as a grid of 4:3 tiles. Tiles stay live (xterm + scrollback + PTY all preserved); bells bounce the tile and outline it; the column-count slider scales the whole grid from "one giant tile" to "10-per-row dense overview." **Single-click** a tile to magnify it to ~70% viewport with keyboard input; **double-click** to enter a dedicated full-pane view that resizes the PTY to fit. **Shift+Cmd+Arrow** (macOS) / **Shift+Ctrl+Arrow** (Linux/Windows) walks to the next-magnified-target tile in the indicated direction — perfect for cycling through a row of build/test pipelines. **Hide / Show Terminals** with named **visibility groupings** (e.g. "claude only", "server logs", "single app") so you can switch contexts without re-toggling rows. A **flow layout mode** flattens projects into a single tagged grid; **sectioned mode** groups by project. Off-screen tiles unmount their xterm to keep memory tame; the PTY stays alive server-side so the bytes are waiting when you scroll back.
+
 **Unread indicators** — when tickets are created or updated externally (by AI tools, sync plugins, or the API), a blue dot appears next to the title. Your own edits in the UI never trigger unread status. Mark as Read/Unread from the context menu or batch toolbar. Tickets are automatically marked as read when you open them in the detail panel.
 
+**Git status tracker** — auto-detected sidebar chip shows your current branch, dirty-file count, and ahead/behind upstream commits. Click to expand a popover with the file lists; one-click `git fetch` updates remote counts without leaving Hot Sheet. Silent no-op for non-git projects.
+
 **Stats dashboard** — click the sidebar widget to open a full analytics page with throughput charts, created-vs-completed trends, cumulative flow diagram, category breakdown, and cycle time scatter plot. Hover any chart for detailed tooltips.
 
 <p align="center">
   <img src="docs/demo-8.png" alt="Stats dashboard showing throughput, flow, and cycle time charts" width="900">
 </p>
 
+<p align="center">
+  <img src="docs/demo-11.png" alt="Embedded terminal in the footer drawer with named tabs for dev server, tests, and Claude" width="900">
+</p>
+
+<p align="center">
+  <img src="docs/demo-12.png" alt="Terminal dashboard view showing every terminal across every registered project as a tile grid" width="900">
+</p>
+
 **Multi-project tabs** — open multiple projects in a single window. Each project remembers its own sidebar view, settings, sort preferences, and layout. Tabs appear automatically when you register a second project via Open Folder (`Cmd+O`). Drag tabs to reorder, right-click for close options and "Show in Finder," switch with `Cmd+Shift+[/]`.
 
 <p align="center">
@@ -124,15 +140,16 @@ The loop stays tight because the AI always knows what to work on next — and yo
 - **Up Next flag** — star tickets to add them to the AI worklist
 - **Drag and drop** — drag tickets onto sidebar views to change category, priority, or status; drop files onto the detail panel to attach; reorder project tabs and custom views
 - **Right-click context menus** — full context menu on tickets with category/priority/status submenus, tags, duplicate, mark as read/unread, backlog, archive, delete
-- **Search** — full-text search across ticket titles, details, ticket numbers, and tags
+- **Search** — full-text search across ticket titles, details, ticket numbers, and tags. When matches sit in the backlog or archive, gray "Include N backlog / N archive" rows surface under the search bar so you can mix them in with one click without leaving the current view
+- **Ticket cross-references** — `HS-NNNN`-style references inside notes, details, and reader-mode become clickable links that open a stacking modal showing the referenced ticket; click another reference inside that modal to push it onto the stack
 - **Print** — print the dashboard, all tickets, selected tickets, or individual tickets in checklist, summary, or full-detail format
 - **Keyboard-driven** — `Enter` to create, `Cmd+I/B/F/R/K/G` for categories, `Alt+1-5` for priority, `Cmd+D` for Up Next, `Delete` to trash, `Cmd+P` to print, `Cmd+Z/Shift+Z` for undo/redo
 - **Undo/redo** — `Cmd+Z` and `Cmd+Shift+Z` for all operations including notes, batch changes, and deletions
 - **Animated transitions** — smooth FLIP animations when tickets reorder after property changes
 - **Copy / cut / paste** — `Cmd+C` copies selected tickets (formatted text to clipboard + structured data for paste), `Cmd+X` cuts, `Cmd+V` pastes into the current project with title dedup. Works across projects.
-- **File attachments** — attach files via file picker or drag-and-drop onto the detail panel, reveal in file manager
+- **File attachments** — attach files via file picker or drag-and-drop onto the detail panel, reveal in file manager. Feedback-dialog drafts also save their attachments — Save Draft persists files alongside the text so you can pick up where you left off.
 - **Markdown sync** — `worklist.md` and `open-tickets.md` auto-generated on every change
-- **Automatic backups** — tiered snapshots (every 5 min, hourly, daily) with preview-before-restore recovery
+- **Automatic backups** — tiered snapshots (every 5 min, hourly, daily) with preview-before-restore recovery. A **Database Repair** panel in Settings can find a working backup automatically and run `pg_resetwal` if the live cluster ever needs it.
 - **Auto-cleanup** — verified tickets auto-archive after a configurable number of days; trashed tickets auto-delete
 - **Portable settings** — all project settings stored in `settings.json` for easy copying between projects
 - **App icon variants** — 9 icon variants to choose from in Settings, applied instantly to the dock icon
@@ -173,10 +190,11 @@ Hot Sheet can push events directly to a running Claude Code session via MCP chan
 - **Play button** — appears in the sidebar. Single-click sends the worklist to Claude on demand. Pending worklist changes are flushed immediately so the AI always reads up-to-date data.
 - **Auto mode** — double-click the play button to enable automatic mode. Claude is triggered immediately, then continues monitoring for new Up Next items with debounce. Exponential backoff prevents runaway retries.
 - **Auto-prioritize** — when no tickets are flagged as Up Next, Claude automatically evaluates open tickets and picks the most important ones to work on.
-- **Feedback loop** — Claude can request user input by adding notes prefixed with `FEEDBACK NEEDED:` or `IMMEDIATE FEEDBACK NEEDED:`. A dialog appears in the UI for the user to respond, and Claude is automatically re-triggered with the feedback. Blue dots on project tabs indicate pending feedback.
-- **Custom commands** — create named buttons that send custom prompts to Claude **or run shell commands** directly. Organize into collapsible groups. Toggle between "Claude Code" and "Shell" targets per command. Shell commands execute server-side with stdout/stderr captured to the commands log.
-- **Permission relay** — when Claude needs tool approval (Bash, Edit, etc.), a full-screen overlay shows the tool name and command preview with Allow/Deny/Dismiss buttons — no need to switch to the terminal.
-- **Commands log** — a resizable bottom panel that records all communication: triggers, completions, permission requests, and shell command output. Filter by type, search, and copy entries. Shell commands show a stop button for running processes.
+- **Feedback loop** — Claude can request user input by adding notes prefixed with `FEEDBACK NEEDED:` or `IMMEDIATE FEEDBACK NEEDED:`. A rich dialog appears in the UI with inline-response slots between every prompt block, a catch-all field, and file attachment support. Save Draft to come back later (text + attachments survive); Submit re-triggers Claude with your response. Blue dots on project tabs indicate pending feedback.
+- **Custom commands** — create named buttons that send custom prompts to Claude **or run shell commands** directly. Organize into collapsible groups. Toggle between "Claude Code" and "Shell" targets per command. Shell commands execute server-side with stdout/stderr streamed to the commands log in real time.
+- **Permission relay** — when Claude needs tool approval (Bash, Edit, etc.), a full-screen overlay shows the tool name and command preview with Allow/Deny/Dismiss buttons — no need to switch to the terminal. Per-project **allow-rules** can auto-approve specific tool+pattern pairs (e.g. always allow `Bash:npm test`) so trusted commands skip the overlay entirely.
+- **Commands log + embedded terminal** — the footer drawer hosts both a resizable commands log (every trigger, completion, permission request, shell output) AND tabs for any number of project-scoped terminals. Hop between an `npm run dev` terminal, the Claude session, and the audit log without leaving Hot Sheet.
+- **MCP tools** — Claude Code sees Hot Sheet as a connected MCP server with schema-validated tools: `hotsheet_create_ticket`, `hotsheet_update_ticket`, `hotsheet_query_tickets`, `hotsheet_batch`, `hotsheet_signal_done`, `hotsheet_request_feedback`, and 8 more. Tickets, notes, and attachments all round-trip without the AI hand-crafting curl commands.
 - **Status indicator** — shows "Claude working" / "Shell running" / idle in the footer.
 
 Requires Claude Code v2.1.80+ with channel support. See [docs/12-claude-channel.md](docs/12-claude-channel.md) for setup details.
@@ -418,8 +436,9 @@ npm install
 
 npm run dev              # Build client assets, then run via tsx
 npm run build            # Build to dist/cli.js
-npm test                 # Unit tests with coverage (626 tests)
-npm run test:e2e         # E2E browser tests (120+ tests)
+npm test                 # Unit tests with coverage (3000+ tests)
+npm run test:e2e         # E2E browser tests (375 tests) — fails on unexpected console errors / pageerrors / 4xx-5xx responses by default (HS-8435 / HS-8436)
+STRICT_E2E_ERRORS=0 npm run test:e2e   # Local debugging only — opt out of the unexpected-event gate
 npm run test:fast        # Unit + fast E2E (skips GitHub plugin tests)
 npm run test:all         # Merged coverage report (unit + E2E)
 npm run lint             # ESLint
@@ -427,7 +446,9 @@ npm run clean            # Remove dist and caches
 npm link                 # Symlink for global 'hotsheet' command
 ```
 
-The project has comprehensive test coverage with 626 unit tests (vitest) and 120+ Playwright E2E browser tests, plus smoke tests for production install verification.
+The project has comprehensive test coverage with 3000+ unit tests (vitest) and 375 Playwright E2E browser tests, plus smoke tests for production install verification.
+
+The E2E suite carries an error-capture fixture (HS-8435 / HS-8436, `e2e/coverage-fixture.ts`) that fails any test with an unexpected `console.error` / `pageerror` / 4xx-5xx response event. Specs that legitimately expect an error opt in via the per-test `errorCapture.allowErrors([patterns])` fixture helper (add a code comment explaining why); global noise lives in the `GLOBAL_ERROR_ALLOWLIST` constant. Set `STRICT_E2E_ERRORS=0` for local debugging only — CI should always run with the strict default.
 
 ---
 

package/dist/channel.js

@@ -34288,7 +34288,12 @@ var DashboardConfigSchema = external_exports3.object({
   layoutMode: external_exports3.enum(["sectioned", "flow"]).optional(),
   columnsPerRow: external_exports3.number().optional(),
   visibilityGroupings: external_exports3.array(VisibilityGroupingSchema).optional(),
-  activeVisibilityGroupingId: external_exports3.string().optional()
+  activeVisibilityGroupingId: external_exports3.string().optional(),
+  // HS-8424 — HS-8406 added per-scope active-grouping selection on the
+  // client; without it here, the `.strict()` parser rejected every
+  // visibility PATCH (the client always sends both keys), so no toggle
+  // made after HS-8406 landed could persist across relaunches.
+  activeVisibilityGroupingIdByScope: external_exports3.record(external_exports3.string(), external_exports3.string()).optional()
 }).strict();
 var GlobalConfigSchema = external_exports3.object({
   channelEnabled: external_exports3.boolean().optional(),
@@ -34369,7 +34374,7 @@ async function proxyRequest(settings, path, init, fetchFn) {
 var UpdateTicketInputSchema = external_exports3.object({
   id: external_exports3.number().int().describe("Ticket id (numeric, e.g. 42)"),
   status: TicketStatusSchema.optional(),
-  notes: external_exports3.string().optional().describe("Replace the ticket's notes JSON. Pass the full notes array as a JSON string."),
+  notes: external_exports3.string().optional().describe('Append a new note. Pass the markdown body as a plain string (NOT a JSON array \u2014 the server wraps the text in `{id, text, created_at}` automatically). HS-8427 \u2014 if you mistakenly pass a JSON-stringified note array like `[{"text":"..."}]`, the server unwraps it defensively so the body renders correctly, but plain text is the documented + preferred shape.'),
   priority: TicketPrioritySchema.optional(),
   category: external_exports3.string().optional().describe('Category id (e.g. "bug", "feature", "task", "issue")'),
   up_next: external_exports3.boolean().optional(),