Add temporary workflow_dispatch debug mode for arXiv fetch diagnosis

Recent runs report 0 papers despite historical volume of 70-150/day. This adds an opt-in diagnostic step to inspect what the GitHub Actions runner actually receives from arxiv.org, without touching digest logic.

Author: richteague

.github/workflows/daily-digest.yml

@@ -13,6 +13,11 @@ on:
     # │ │  │ │ │
     - cron: "0 12 * * 1-5"
   workflow_dispatch: # Also allows manual runs from the GitHub Actions tab
+    inputs:
+      debug_fetch:
+        description: "Diagnostic only: fetch arxiv.org/list/astro-ph/new and print status/length instead of running the digest"
+        type: boolean
+        default: false
 
 permissions:
   contents: write # to commit posted-ids.json back after each run
@@ -34,7 +39,32 @@ jobs:
       - name: Install dependencies
         run: npm ci
 
+      - name: Debug arXiv fetch
+        if: ${{ inputs.debug_fetch }}
+        run: |
+          node -e '
+          fetch("https://arxiv.org/list/astro-ph/new").then(async res => {
+            console.log("status:", res.status);
+            for (const [k, v] of res.headers.entries()) console.log(k + ":", v);
+            const html = await res.text();
+            console.log("html length:", html.length);
+            const now = new Date();
+            const MONTHS = ["January","February","March","April","May","June","July","August","September","October","November","December"];
+            const todayStr = `${now.getUTCDate()} ${MONTHS[now.getUTCMonth()]} ${now.getUTCFullYear()}`;
+            console.log("todayStr:", todayStr, "included:", html.includes(todayStr));
+            const newSection = html.split(/Replacement submissions/i)[0];
+            console.log("newSection length:", newSection.length);
+            const ids = [...new Set([...newSection.matchAll(/arXiv:(\d{4}\.\d{4,5})/g)].map(m => m[1]))];
+            console.log("ids found:", ids.length);
+            const h3 = [...html.matchAll(/<h3>([^<]*)<\/h3>/g)].map(m => m[1]);
+            console.log("h3 headings:", JSON.stringify(h3));
+            console.log("--- first 500 chars ---");
+            console.log(html.slice(0, 500));
+          }).catch(err => { console.error("fetch failed:", err); process.exit(1); });
+          '
+
       - name: Run Disk Digest
+        if: ${{ !inputs.debug_fetch }}
         env:
           SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
           SLACK_CHANNEL_ID: ${{ secrets.SLACK_CHANNEL_ID }}
@@ -44,6 +74,7 @@ jobs:
 
       # Persist the record of already-digested papers so later runs skip them
       - name: Commit posted-ids.json
+        if: ${{ !inputs.debug_fetch }}
         run: |
           test -f posted-ids.json || exit 0
           git config user.name "github-actions[bot]"