diff --git a/.github/workflows/audit_poc_links.yml b/.github/workflows/audit_poc_links.yml
index d817fad4f61..011aa323188 100644
--- a/.github/workflows/audit_poc_links.yml
+++ b/.github/workflows/audit_poc_links.yml
@@ -16,7 +16,7 @@ concurrency:
 jobs:
   audit:
     runs-on: ubuntu-latest
-    timeout-minutes: 30
+    timeout-minutes: 50
     steps:
       - name: Checkout repository
         uses: actions/checkout@v7
@@ -26,14 +26,13 @@ jobs:
         with:
           python-version: '3.13'
 
-      # A seventh of the old weekly batch every day: the 12,000-and-12,000
-      # weekly run measured an hour, and it holds the repository-updates lock
-      # the whole time, which stalls the hourly refresh and can cancel a daily
-      # sync queued behind it. Same sweep period, about nine minutes a day.
+      # 4,000 repositories and 3,000 reference links a day sweeps the whole
+      # index about weekly, in around half an hour. A sample found one link in
+      # eight pointing at a deleted repository under the old fortnightly sweep.
       - name: Drop dead links and refresh repository metadata
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: python3 scripts/audit_poc_links.py --limit 1750 --references 1750
+        run: python3 scripts/audit_poc_links.py --limit 4000 --references 3000
 
       - name: Verify the generated index
         run: python3 scripts/build_site.py
diff --git a/.github/workflows/hot_cves.yml b/.github/workflows/hot_cves.yml
index e9d9bf4620b..a177b3495f1 100644
--- a/.github/workflows/hot_cves.yml
+++ b/.github/workflows/hot_cves.yml
@@ -26,6 +26,15 @@ jobs:
         with:
           python-version: '3.13'
 
+      # Every repository pushed in the last day that names a CVE of this year
+      # or last, with its README read and the classifier applied, so a PoC
+      # published this morning is on the site within the hour. The daily run
+      # keeps covering every year and the CVE record references.
+      - name: Discover this year's PoCs pushed in the last day
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: python3 scripts/update_cves.py --lookback-days 1 --years 2 --skip-cvelist
+
       - name: Ingest the previous hour's candidates
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -52,6 +61,8 @@ jobs:
             README.md \
             index/trending.json \
             index/github.txt \
+            index/references.txt \
+            index/reference_pocs.txt \
             index/cve_dates.json \
             index/cve_record_state.json \
             docs/hero.svg \
diff --git a/.github/workflows/sync_pocingithub.yml b/.github/workflows/sync_pocingithub.yml
index d25dada0fba..a40bff5d54e 100644
--- a/.github/workflows/sync_pocingithub.yml
+++ b/.github/workflows/sync_pocingithub.yml
@@ -2,7 +2,7 @@ name: Sync PoC-in-GitHub
 
 on:
   schedule:
-    - cron: '10 5 * * 3'
+    - cron: '10 5 * * *'
   workflow_dispatch:
 
 permissions:
@@ -26,10 +26,11 @@ jobs:
         with:
           python-version: '3.13'
 
-      # Every candidate is judged again each week rather than remembered as
+      # Every candidate is judged again each day rather than remembered as
       # rejected: a repository that was an empty shell in March is often a
       # working exploit by June, and re-reading 6,800 of them costs five
-      # minutes.
+      # minutes. Daily because the collection keeps finding repositories our
+      # own search missed.
       - name: Judge upstream PoC candidates
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
diff --git a/scripts/audit_poc_links.py b/scripts/audit_poc_links.py
index 48d1472a88a..340c036110b 100644
--- a/scripts/audit_poc_links.py
+++ b/scripts/audit_poc_links.py
@@ -16,6 +16,8 @@ from urllib.parse import urlsplit
 sys.path.insert(0, str(Path(__file__).resolve().parent))
 
 from update_cves import (
+    readme_text,
+    repo_is_bare,
     GITHUB_GRAPHQL_URL,
     kev_badge,
     USER_AGENT,
@@ -72,11 +74,24 @@ def collect_repositories() -> tuple[dict[str, list[tuple[Path, str]]], set[str]]
     return prunable, referenced
 
 
-def survey(client: GitHubClient, names: list[str]) -> tuple[set[str], dict[str, list]]:
-    """Return the repositories GitHub reports as NOT_FOUND, and stars/last push for the rest.
+REPO_FIELDS = (
+    "{ nameWithOwner stargazerCount pushedAt "
+    'readmeMd: object(expression: "HEAD:README.md") { ... on Blob { text } } '
+    'readmeUpper: object(expression: "HEAD:README.MD") { ... on Blob { text } } '
+    'readmeRst: object(expression: "HEAD:README.rst") { ... on Blob { text } } '
+    'readmeBare: object(expression: "HEAD:README") { ... on Blob { text } } '
+    'root: object(expression: "HEAD:") { ... on Tree { entries { name type } } } }'
+)
 
-    Anything else, a network error or a throttle or an unexpected response, leaves
-    the repository out of both results, so an audit failure never drops a link.
+
+def survey(client: GitHubClient, names: list[str]) -> tuple[set[str], dict[str, list]]:
+    """Return the repositories to drop, and stars/last push for the rest.
+
+    Dropped means GitHub reports NOT_FOUND, or the repository is bare: no code,
+    no directory, no artifact and a README with nothing runnable in it, which
+    the discovery gate would reject today. Anything else, a network error or a
+    throttle or an unexpected response, leaves the repository out of both
+    results, so an audit failure never drops a link.
     """
     batches = [names[index : index + BATCH] for index in range(0, len(names), BATCH)]
     missing: set[str] = set()
@@ -87,7 +102,7 @@ def survey(client: GitHubClient, names: list[str]) -> tuple[set[str], dict[str,
         aliases = " ".join(
             f"r{index}: repository(owner: {json.dumps(name.split('/', 1)[0])}, "
             f"name: {json.dumps(name.split('/', 1)[1])}) "
-            "{ nameWithOwner stargazerCount pushedAt }"
+            f"{REPO_FIELDS}"
             for index, name in enumerate(batch)
         )
         payload = http_json(
@@ -108,6 +123,9 @@ def survey(client: GitHubClient, names: list[str]) -> tuple[set[str], dict[str,
             repo = data.get(f"r{index}")
             if not repo:
                 continue
+            if repo_is_bare(repo, readme_text(repo)):
+                gone.add(name)
+                continue
             pushed = str(repo.get("pushedAt") or "")[:10]
             found[name.lower()] = [
                 int(repo.get("stargazerCount") or 0),
@@ -418,7 +436,7 @@ def main() -> int:
         GitHubClient(token),
         [name for name in scope if name not in blocked],
     )
-    print(f"{len(missing)} repositories no longer exist")
+    print(f"{len(missing)} repositories no longer exist or hold nothing runnable")
     if blocked:
         print(f"Pruning {len(blocked)} explicitly blacklisted repositories")
     missing |= blocked
diff --git a/scripts/update_cves.py b/scripts/update_cves.py
index 1d8f3a80e84..3fde86c48b9 100755
--- a/scripts/update_cves.py
+++ b/scripts/update_cves.py
@@ -291,6 +291,41 @@ def root_has_code(repo: dict[str, Any]) -> bool:
     return any(CODE_FILE_RE.search(name) for name in root_entry_names(repo))
 
 
+TEXT_ONLY_RE = re.compile(
+    r"^(?:readme|license|licence|security|contributing|changelog|code_of_conduct|"
+    r"\.gitignore|\.gitattributes|funding\.yml)(?:\.|$)|\.(?:md|txt|rst)$",
+    re.IGNORECASE,
+)
+PAYLOAD_RE = re.compile(r"```|~~~|^ {4}\S", re.MULTILINE)
+COMMAND_RE = re.compile(
+    r"\b(?:curl|wget|python3?|php|perl|ruby|nc|ncat|bash|msfconsole|nuclei|sqlmap|"
+    r"GET /|POST /|PUT /)\b|https?://[^\s)]*\?"
+)
+
+
+def repo_is_bare(repo: dict[str, Any], readme: str) -> bool:
+    """Nothing runnable, and nothing that says how.
+
+    A sample of two hundred indexed repositories held fourteen with a bare
+    README or none at all: empty shells, placeholders, a link to a paywall.
+    Anything with a directory, a file that is not documentation, a file named
+    for the CVE or as a PoC, or a README that carries a code block, a command,
+    a request or a real write-up is not bare. Where the root is unknown the
+    question is not asked.
+    """
+    root = repo.get("root")
+    if not isinstance(root, dict):
+        return False
+    for entry in root.get("entries") or []:
+        name = str(entry.get("name") or "")
+        if str(entry.get("type") or "") == "tree" or not TEXT_ONLY_RE.search(name):
+            return False
+        if POC_RE.search(normalise_identity(name)) or CODE_FILE_RE.search(name):
+            return False
+    text = readme or ""
+    return not (len(text) >= 1000 or PAYLOAD_RE.search(text) or COMMAND_RE.search(text))
+
+
 def truncated_ids(cve_ids: Collection[str]) -> set[str]:
     """Identifiers that a longer identifier from the same year extends."""
     numbers: dict[str, set[str]] = defaultdict(set)
@@ -343,6 +378,8 @@ def qualifying_repo_cves(
     description = str(repo.get("description") or "")
     topics = " ".join(topic_names(repo))
     readme = readme_text(repo)
+    if repo_is_bare(repo, readme):
+        return set()
     all_name_cves = extract_cves(full_name) | extract_cves(normalise_identity(full_name))
     all_identity_cves = (
         all_name_cves