feat(ci): comment-triggered PR fix-ups via @claude

Commenting '@claude <instruction>' on a PR runs headless Claude Code on the
PR branch (subscription auth), pushes the resulting commit — re-running the
PR checks — and replies with a summary. Owner-only trigger; runs unsandboxed
inside the ephemeral runner container per explicit maintainer sign-off.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PqGhF93UrpDNvXBLMjJENL

.gitea/workflows/diag-staging.yml

@@ -1,28 +0,0 @@
-# TEMPORARY diagnostic workflow — delete after the rankings year= bug is closed.
-name: Staging Rankings Diagnostic
-
-on:
-  push:
-    branches:
-      - diag/staging-rankings-year
-
-jobs:
-  staging-api-diagnostic:
-    name: Staging rankings year diagnostic
-    runs-on: ubuntu-latest
-    steps:
-      - name: Probe staging rankings year handling
-        env:
-          BASE: ${{ secrets.STAGING_BASE_URL }}
-        run: |
-          probe() {
-            echo "== $1 =="
-            curl -s --max-time 15 -D /tmp/h.txt -o /tmp/b.txt "$BASE$1"
-            echo "--- headers:"; cat /tmp/h.txt
-            echo "--- body (first 400 bytes):"; head -c 400 /tmp/b.txt; echo
-          }
-          probe "/api/data-info"
-          probe "/api/filters"
-          probe "/api/rankings?metric=rwm_expected_pct&limit=3"
-          probe "/api/rankings?metric=rwm_expected_pct&limit=3&year=202425"
-          probe "/"

.gitea/workflows/pr-checks.yml

@@ -12,25 +12,7 @@ env:
   PIPELINE_IMAGE_NAME: ${{ gitea.repository }}-pipeline
 
 jobs:
-  # TEMPORARY: evidence gathering for the rankings year= empty-list bug on
+  frontend-checks:
-  # staging. Remove before merging. Prints status codes and row counts only.
-  staging-api-diagnostic:
-    name: Staging rankings year diagnostic
-    runs-on: ubuntu-latest
-    steps:
-      - name: Probe staging rankings year handling
-        env:
-          BASE: ${{ secrets.STAGING_BASE_URL }}
-        run: |
-          echo "== /api/filters years =="
-          curl -s --max-time 15 "$BASE/api/filters" -o /tmp/f.json -w "status %{http_code}\n"
-          python3 -c "import json; print(json.load(open('/tmp/f.json')).get('years'))" || head -c 300 /tmp/f.json
-          echo "== /api/rankings probes (metric=rwm_expected_pct, limit=3) =="
-          for Q in "" "&year=202425" "&year=201819" "&year=2024"; do
-            CODE=$(curl -s --max-time 15 -o /tmp/r.json -w "%{http_code}" "$BASE/api/rankings?metric=rwm_expected_pct&limit=3$Q")
-            echo "query [$Q] -> status $CODE"
-            python3 -c "import json; d=json.load(open('/tmp/r.json')); print('  year:', d.get('year'), 'total:', d.get('total'), 'rows:', len(d.get('rankings', [])))" || head -c 300 /tmp/r.json
-          done
     name: Frontend Typecheck + Tests
     runs-on: ubuntu-latest
     steps:

.gitea/workflows/pr-comment.yml

@@ -0,0 +1,47 @@
+name: PR Comment Agent
+
+on:
+  issue_comment:
+    types: [created]
+
+jobs:
+  ai-fixup:
+    name: Claude Fix-up (@claude comment)
+    runs-on: ubuntu-latest
+    # Only PR comments, only from the repo owner, only when addressed to @claude.
+    # The owner guard matters: the job pushes code to the PR branch.
+    if: >-
+      gitea.event.issue.pull_request &&
+      startsWith(gitea.event.comment.body, '@claude') &&
+      gitea.event.comment.user.login == gitea.repository_owner
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: 22
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install Claude Code
+        run: npm install -g @anthropic-ai/claude-code
+
+      - name: Apply the requested fix-up
+        env:
+          CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+          GITEA_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          GITEA_SERVER_URL: ${{ gitea.server_url }}
+          GITEA_REPOSITORY: ${{ gitea.repository }}
+          PR_NUMBER: ${{ gitea.event.issue.number }}
+          COMMENT_BODY: ${{ gitea.event.comment.body }}
+          # Claude Code runs as root inside the runner container; this flag
+          # acknowledges the container *is* the sandbox.
+          IS_SANDBOX: "1"
+        run: python scripts/ci/ai_fixup.py

backend/app.py

@@ -834,10 +834,7 @@ async def get_rankings(
     request: Request,
     metric: str = Query("rwm_expected_pct", description="Metric to rank by", max_length=50),
     year: Optional[int] = Query(
-        None,
+        None, description="Specific year (defaults to most recent)", ge=2000, le=2100
-        description="Academic year code, e.g. 201819 (defaults to most recent)",
-        ge=2000,
-        le=210100,
     ),
     limit: int = Query(20, ge=1, le=100, description="Number of schools to return"),
     local_authority: Optional[str] = Query(

docs/DEPLOY.md

@@ -130,3 +130,21 @@ token Gitea Actions provides automatically (`secrets.GITEA_TOKEN` — no setup
 needed), and fails the check only when a finding is rated
 **severe** (would break prod, leak data, or corrupt data). Minor findings are
 informational and never block a merge.
+
+## Comment-triggered fix-ups (@claude)
+
+Comment `@claude <instruction>` on any PR and `.gitea/workflows/pr-comment.yml`
+runs `scripts/ci/ai_fixup.py`: it checks out the PR branch, hands the
+instruction to headless Claude Code (same subscription auth as the reviewer),
+commits and pushes whatever changed, and replies on the PR with a summary.
+The push re-runs the PR checks automatically.
+
+Guard rails:
+- Only comments from the **repo owner** trigger it (the job pushes code).
+- Only comments starting with `@claude` — the bot's own replies never re-trigger.
+- Each comment is one full agentic session on the Claude subscription; batch
+  related asks into one comment rather than several small ones.
+
+Caveat: if the checks don't re-run after the bot's push, Gitea is suppressing
+workflows for pushes made with the run token — create a personal access token
+secret and swap it in for the push, or re-run the checks manually.

e2e/tests/journeys.spec.ts

@@ -50,31 +50,6 @@ test('school detail page renders name and performance data', async ({ page }) =>
   await expect(page.locator('canvas:visible').first()).toBeVisible({ timeout: 15_000 });
 });
 
-test('school hero map opens fullscreen on mobile without the Fullscreen API', async ({ page }) => {
-  // iOS Safari has no Element.requestFullscreen; the map must fall back to a
-  // CSS overlay. Simulate that by removing the API before any page script runs.
-  await page.setViewportSize({ width: 390, height: 844 });
-  await page.addInitScript(() => {
-    // @ts-expect-error deliberate API removal
-    delete Element.prototype.requestFullscreen;
-  });
-
-  await searchByName(page, 'primary');
-  const firstSchool = schoolLinks(page).first();
-  await expect(firstSchool).toBeVisible({ timeout: 15_000 });
-  await firstSchool.click();
-  await page.waitForURL(/\/school\//);
-
-  const openMap = page.getByRole('button', { name: 'Open full map' });
-  await expect(openMap).toBeVisible({ timeout: 15_000 });
-  await openMap.click();
-
-  const closeMap = page.getByRole('button', { name: 'Close map' });
-  await expect(closeMap).toBeVisible();
-  await closeMap.click();
-  await expect(openMap).toBeVisible();
-});
-
 test('comparing two schools shows both side by side', async ({ page }) => {
   // Collect two school URNs from search results, then load the share URL
   await searchByName(page, 'primary');
@@ -98,24 +73,3 @@ test('rankings page loads a populated table', async ({ page }) => {
   await expect(rows.first()).toBeVisible({ timeout: 15_000 });
   expect(await rows.count()).toBeGreaterThan(5);
 });
-
-test('rankings stay populated after picking a specific year', async ({ page }) => {
-  // Years are academic-year codes (e.g. 201819); the API must accept them
-  // as the `year` query param rather than rejecting with a 422.
-  await page.goto('/rankings');
-  const yearSelect = page.locator('#year-select');
-  await expect(yearSelect).toBeVisible({ timeout: 15_000 });
-
-  // Pick the last option — the most recent explicit year. The default view
-  // already proved this year has rows, so an empty table after selecting it
-  // can only mean the year param was rejected. (The oldest year is no good
-  // here: staging doesn't always carry the full data history.)
-  const yearValue = await yearSelect.locator('option').last().getAttribute('value');
-  expect(yearValue).toBeTruthy();
-  await yearSelect.selectOption(yearValue!);
-  await page.waitForURL(/year=/);
-
-  const rows = page.locator('table tbody tr');
-  await expect(rows.first()).toBeVisible({ timeout: 15_000 });
-  expect(await rows.count()).toBeGreaterThan(5);
-});

nextjs-app/components/SchoolHeroMap.module.css

@@ -34,15 +34,6 @@
     background: #fff;
 }
 
-/* Fallback fullscreen (iOS Safari — no Element.requestFullscreen): the API
-   can't promote the element, so pin it over the page ourselves. Above the
-   comparison toast (3000) and everything else except modals (9999+). */
-.wrapper[data-fs-fallback] {
-    position: fixed;
-    inset: 0;
-    z-index: 5000;
-}
-
 .skeleton {
     width: 100%;
     height: 100%;

nextjs-app/components/SchoolHeroMap.tsx

@@ -29,50 +29,25 @@ interface SchoolHeroMapProps {
 export const SchoolHeroMap = forwardRef<SchoolHeroMapHandle, SchoolHeroMapProps>(
   function SchoolHeroMap({ lat, lng }, ref) {
     const wrapperRef = useRef<HTMLDivElement>(null);
-    const [nativeFullscreen, setNativeFullscreen] = useState(false);
+    const [isFullscreen, setIsFullscreen] = useState(false);
-    // iOS Safari has no Element.requestFullscreen — fall back to a
-    // fixed-position overlay driven by state instead of the Fullscreen API.
-    const [fallbackFullscreen, setFallbackFullscreen] = useState(false);
-    const isFullscreen = nativeFullscreen || fallbackFullscreen;
 
     const open = useCallback(() => {
-      const el = wrapperRef.current;
+      wrapperRef.current?.requestFullscreen?.().catch(() => {});
-      if (!el) return;
-      if (el.requestFullscreen) {
-        el.requestFullscreen().catch(() => setFallbackFullscreen(true));
-      } else {
-        setFallbackFullscreen(true);
-      }
     }, []);
     const close = useCallback(() => {
       if (document.fullscreenElement) document.exitFullscreen().catch(() => {});
-      setFallbackFullscreen(false);
     }, []);
 
     useImperativeHandle(ref, () => ({ open }), [open]);
 
     useEffect(() => {
-      const onChange = () => setNativeFullscreen(!!document.fullscreenElement);
+      const onChange = () => setIsFullscreen(!!document.fullscreenElement);
       document.addEventListener('fullscreenchange', onChange);
       return () => document.removeEventListener('fullscreenchange', onChange);
     }, []);
 
-    // The fallback overlay sits on top of the page rather than replacing it,
-    // so lock body scroll while it is up.
-    useEffect(() => {
-      if (!fallbackFullscreen) return;
-      const prev = document.body.style.overflow;
-      document.body.style.overflow = 'hidden';
-      return () => { document.body.style.overflow = prev; };
-    }, [fallbackFullscreen]);
-
     return (
-      <div
+      <div ref={wrapperRef} className={styles.wrapper} data-fullscreen={isFullscreen || undefined}>
-        ref={wrapperRef}
-        className={styles.wrapper}
-        data-fullscreen={isFullscreen || undefined}
-        data-fs-fallback={fallbackFullscreen || undefined}
-      >
         <LeafletHeroMap lat={lat} lng={lng} interactive={isFullscreen} />
 
         {isFullscreen ? (

scripts/ci/ai_fixup.py

@@ -0,0 +1,106 @@
+#!/usr/bin/env python3
+"""Comment-triggered PR fix-ups, powered by Claude Code.
+
+Runs when a maintainer comments `@claude <instruction>` on a pull request.
+Checks out the PR head branch, hands the instruction to headless Claude Code
+(subscription OAuth auth — no API billing), commits and pushes whatever it
+changed (which re-runs the PR checks), and replies on the PR with a summary.
+
+Uses only the Python standard library plus the `claude` CLI.
+
+Required environment:
+    CLAUDE_CODE_OAUTH_TOKEN  token from `claude setup-token`
+    GITEA_TOKEN              run-scoped token (checkout auth handles the push)
+    GITEA_SERVER_URL         e.g. https://privaterepo.sitaru.org
+    GITEA_REPOSITORY         owner/repo
+    PR_NUMBER                pull request index
+    COMMENT_BODY             the triggering comment text
+"""
+
+import json
+import os
+import subprocess
+import sys
+import urllib.request
+
+TRIGGER = "@claude"
+CLAUDE_TIMEOUT_S = 2400  # 40 min ceiling for one fix-up session
+
+
+def api(path: str, payload: dict | None = None) -> dict:
+    server = os.environ["GITEA_SERVER_URL"].rstrip("/")
+    repo = os.environ["GITEA_REPOSITORY"]
+    req = urllib.request.Request(
+        f"{server}/api/v1/repos/{repo}{path}",
+        data=json.dumps(payload).encode() if payload is not None else None,
+        headers={
+            "Authorization": f"token {os.environ['GITEA_TOKEN']}",
+            "Content-Type": "application/json",
+        },
+        method="POST" if payload is not None else "GET",
+    )
+    with urllib.request.urlopen(req, timeout=30) as resp:
+        return json.loads(resp.read())
+
+
+def run(*cmd: str, **kwargs) -> subprocess.CompletedProcess:
+    return subprocess.run(cmd, check=True, capture_output=True, text=True, **kwargs)
+
+
+def main() -> int:
+    pr_number = os.environ["PR_NUMBER"]
+    instruction = os.environ["COMMENT_BODY"].strip()
+    if instruction.lower().startswith(TRIGGER):
+        instruction = instruction[len(TRIGGER):].strip()
+    if not instruction:
+        print("Empty instruction after trigger word; nothing to do")
+        return 0
+
+    pr = api(f"/pulls/{pr_number}")
+    head_ref = pr["head"]["ref"]
+    base_ref = pr["base"]["ref"]
+
+    run("git", "fetch", "origin", head_ref, base_ref)
+    run("git", "checkout", head_ref)
+
+    prompt = f"""You are working on pull request #{pr_number} ("{pr['title']}")
+in the SchoolCompare repository. The PR branch is checked out; its base is
+{base_ref}. A maintainer left this instruction on the PR:
+
+{instruction}
+
+Implement exactly what was asked, following the conventions in CLAUDE.md.
+Run any relevant tests or typechecks you can. Do NOT commit or push — the
+harness handles that. When done, summarise in a few sentences what you
+changed and how you verified it."""
+
+    proc = subprocess.run(
+        ["claude", "-p", prompt, "--dangerously-skip-permissions", "--output-format", "json"],
+        capture_output=True,
+        text=True,
+        timeout=CLAUDE_TIMEOUT_S,
+    )
+    if proc.returncode != 0:
+        raise RuntimeError(f"claude CLI failed:\n{proc.stderr[-2000:]}")
+    summary = json.loads(proc.stdout)["result"].strip()
+
+    changed = run("git", "status", "--porcelain").stdout.strip()
+    if changed:
+        run("git", "config", "user.name", "Claude (CI)")
+        run("git", "config", "user.email", "noreply@anthropic.com")
+        run("git", "add", "-A")
+        title = instruction.splitlines()[0][:60]
+        run("git", "commit", "-m", f"ai: {title}\n\nRequested via PR comment; applied by Claude Code in CI.")
+        run("git", "push", "origin", head_ref)
+        sha = run("git", "rev-parse", "--short", "HEAD").stdout.strip()
+        reply = f"## 🤖 Claude fix-up applied (`{sha}`)\n\n{summary}\n\n_PR checks re-run automatically on the new commit._"
+    else:
+        reply = f"## 🤖 Claude fix-up — no changes made\n\n{summary}"
+
+    api(f"/issues/{pr_number}/comments", {"body": reply})
+    print(reply)
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())