## Implementation Plan: Image Paste into Issue / Comment Textareas

### Objective
Add the ability for users to paste images (Ctrl+V / Cmd+V) directly from the clipboard into the issue description textarea and all comment textareas. The image is uploaded via the existing `POST /upload` endpoint and inserted as markdown `![](url)` at the cursor position.

### Files to Modify

#### 1. `static/js/IssueReporter.js`
- **Add a new method** `_onPaste(event)` to the `IssueReporter` class (or as a module function).
- **Logic**:
  ```javascript
  _onPaste(event) {
    const items = event.clipboardData?.items;
    if (!items) return;
    for (const item of items) {
      if (item.kind === 'file' && item.type.startsWith('image/')) {
        event.preventDefault(); // stop default paste (plain text)
        const file = item.getAsFile();
        this._uploadImage(file, event.target);
        break;
      }
    }
  }

  _uploadImage(file, textarea) {
    // 1. Show a small loading indicator (optional) inside the textarea or as a toast.
    // 2. Use `Http.sendForm()` or `fetch()` to POST to /upload
    //    Form data: `file` with key from server (likely "file")
    // 3. On success, get `{ uid, url, ... }`.
    // 4. Call TextInput.insertAtCursor(textarea, `![](url)`).
    // 5. Dispatch `input` event on textarea (TextInput already does this).
    // 6. On error, show error toast.
  }
  ```
- **Where to bind**: In the `init()` or constructor, after the textarea element is known:
  ```javascript
  this.textarea?.addEventListener('paste', this._onPaste.bind(this));
  ```
- **Textarea selector**: `document.querySelector('#issue-description')` – already present in the file.

#### 2. `static/js/CommentManager.js`
- **Same pattern** as `IssueReporter.js`. Add a `_onPaste` method and bind it to the comment textareas.
- **Textarea selectors**:
  - For issue detail page: `textarea[name='body']` (inside `#issue-detail`).
  - For site-wide comment form: `textarea[name='content']` (inside `._comment-form`).
- **Binding location**: In the method that initialises the comment form (e.g., `init()` or after AJAX load). Iterate over all matching textareas and attach the listener.

#### 3. `static/js/TextInput.js` – No changes needed
The existing `TextInput.insertAtCursor(element, text)` handles cursor position restoration and `input` event dispatch. We will reuse it.

#### 4. Backend (no changes)
- `routers/uploads.py` – already accepts `POST /upload` with a `file` field.
- `models.py` – `IssueForm` and `IssueCommentForm` already have `attachment_uids`.
- `routers/issues/attachments.py` – linking logic intact.

#### 5. Templates (no changes)
All target textareas already have stable class/id and name attributes.

### Testing (New Tests Required)
Because no paste tests exist, we will add a Python integration test that simulates a paste-like upload (multipart file upload) and verifies the returned UID is usable. This is a minimal addition that exercises the upload pathway that will be triggered by paste.

- **File**: `tests/api/uploads/test_paste_upload.py`
- **Content**: Upload an image file via HTTP POST to `/upload` with the same form key used by the frontend. Verify response contains `uid` and `url`. This test already exists in `tests/api/uploads/upload.py` but we add a specific test name to document paste coverage.
- Also ensure the existing `tests/api/issues/create.py` and `tests/api/issues/attachments.py` pass unchanged – they already test attachment linking.

### Implementation Steps (Agent Execution Order)

1. Add paste handler to `static/js/IssueReporter.js` (about 25 lines).
2. Add paste handler to `static/js/CommentManager.js` (about 30 lines, allowing for multiple textareas).
3. Add a new test file `tests/api/uploads/test_paste_upload.py` (or append to `upload.py`) with a simple test that uploads a PNG and verifies response shape.
4. Run `pip install -e '.[dev]' -q && make test-unit` – verify all tests pass.
5. Run `pip install -q ruff && ruff check .` – no new lint errors.
6. Commit all changes with a descriptive message.

### Definition of Done
- [ ] All changes are committed to the repository and `git diff --stat` shows only the intended file modifications.
- [ ] The project's test command passes:
  ```
  pip install -e '.[dev]' -q && make test-unit
  ```
  (All existing Python unit tests pass; the new upload test also passes.)
- [ ] The project's lint command passes:
  ```
  pip install -q ruff && ruff check .
  ```
  (No Python lint errors.)
- [ ] Manual/QA verification (for a human after deployment): Pasting a PNG/JPG from clipboard into the issue description textarea inserts `![](/uploads/...)` at cursor position and the rendered image appears after save.
- [ ] Same behaviour confirmed for comment textareas on issue detail page and site-wide comment form.

No backend or template changes were required, and the total new JavaScript code is under 60 lines.