Filter trending PoCs to current-year updates

.github/getTrending.py

@@ -1,99 +1,114 @@
 #!/usr/bin/env python3
-# -*- coding:utf-8 -*-
+"""Regenerate the Trending PoCs table in README.md.
+
+- Only consider repositories whose names contain the current year's CVE pattern (e.g., CVE-2025-1234).
+- Restrict to repositories updated in the last 4 days.
+- Sort by most recently updated, then stars, and emit up to 20 rows.
+"""
+
+from __future__ import annotations
+
+import os
+import re
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+from typing import Iterable, List, TypedDict
+
 import requests
-import json
-from datetime import datetime
+
+WINDOW_DAYS = 4
+MAX_ROWS = 20
 
 
-def time_ago(datetime_str):
-    datetime_obj = datetime.strptime(datetime_str, "%Y-%m-%dT%H:%M:%SZ")
-    current_datetime = datetime.now()
-    delta = current_datetime - datetime_obj
+class Repo(TypedDict):
+    name: str
+    html_url: str
+    description: str | None
+    stargazers_count: int
+    updated_at: str
+
+
+def github_headers() -> dict:
+    token = os.environ.get("GITHUB_TOKEN") or os.environ.get("GH_TOKEN")
+    headers = {"Accept": "application/vnd.github+json"}
+    if token:
+        headers["Authorization"] = f"Bearer {token}"
+    return headers
+
+
+def time_ago(updated_at: str, now: datetime) -> str:
+    dt = datetime.strptime(updated_at, "%Y-%m-%dT%H:%M:%SZ").replace(tzinfo=timezone.utc)
+    delta = now - dt
     if delta.days > 0:
-        if delta.days == 1:
-            return "1 day ago"
-        else:
-            return f"{delta.days} days ago"
-    elif delta.seconds >= 3600:
-        hours = delta.seconds // 3600
-        if hours == 1:
-            return "1 hour ago"
-        else:
-            return f"{hours} hours ago"
-    elif delta.seconds >= 60:
-        minutes = delta.seconds // 60
-        if minutes == 1:
-            return "1 minute ago"
-        else:
-            return f"{minutes} minutes ago"
+        return "1 day ago" if delta.days == 1 else f"{delta.days} days ago"
+    hours = delta.seconds // 3600
+    if hours:
+        return "1 hour ago" if hours == 1 else f"{hours} hours ago"
+    minutes = (delta.seconds % 3600) // 60
+    if minutes:
+        return "1 minute ago" if minutes == 1 else f"{minutes} minutes ago"
+    return "just now"
+
+
+def fetch_trending(current_year: int, cutoff: datetime) -> List[Repo]:
+    query = f"CVE-{current_year} in:name stars:>2 pushed:>={cutoff.date().isoformat()} archived:false"
+    url = "https://api.github.com/search/repositories"
+    params = {
+        "q": query,
+        "sort": "updated",
+        "order": "desc",
+        "per_page": 100,
+        "page": 1,
+    }
+    resp = requests.get(url, params=params, headers=github_headers(), timeout=30)
+    resp.raise_for_status()
+    items: Iterable[Repo] = resp.json().get("items", [])
+    pattern = re.compile(rf"cve-{current_year}-\d+", re.IGNORECASE)
+    filtered: List[Repo] = []
+    for item in items:
+        name = item.get("name", "")
+        updated_at = item.get("updated_at")
+        if not updated_at or not pattern.search(name or ""):
+            continue
+        updated_dt = datetime.strptime(updated_at, "%Y-%m-%dT%H:%M:%SZ").replace(tzinfo=timezone.utc)
+        if updated_dt < cutoff:
+            continue
+        filtered.append(item)
+    # Already sorted by updated desc; break ties by stars
+    filtered.sort(key=lambda r: (-datetime.strptime(r["updated_at"], "%Y-%m-%dT%H:%M:%SZ").timestamp(), -int(r.get("stargazers_count", 0))))
+    return filtered[:MAX_ROWS]
+
+
+def build_rows(repos: List[Repo], now: datetime) -> List[str]:
+    rows: List[str] = []
+    for repo in repos:
+        desc = repo.get("description") or ""
+        stars = int(repo.get("stargazers_count", 0))
+        updated = time_ago(repo["updated_at"], now)
+        rows.append(f"| {stars}⭐ | {updated} | [{repo['name']}]({repo['html_url']}) | {desc} |")
+    return rows
+
+
+def main() -> None:
+    current_year = datetime.now(timezone.utc).year
+    cutoff = datetime.now(timezone.utc) - timedelta(days=WINDOW_DAYS)
+    now = datetime.now(timezone.utc)
+
+    repos = fetch_trending(current_year, cutoff)
+
+    output: List[str] = ['<h1 align="center">Recently updated Proof-of-Concepts</h1>']
+    output.append(f"\n\n## {current_year}\n")
+    output.append(f"### Updated in the last {WINDOW_DAYS} days (up to {MAX_ROWS} repos)\n")
+    output.append("| Stars | Updated | Name | Description |")
+    output.append("| --- | --- | --- | --- |")
+    if repos:
+        output.extend(build_rows(repos, now))
     else:
-        return "just now"
+        output.append("| 0⭐ | — | No recent CVE PoCs | No repositories matched the filters. |")
 
-current_year = datetime.now().year
-total_repos_per_year = {}
-#tz_header = {"Time-Zone": "Europe/Amsterdam"}
+    Path("README.md").write_text("\n".join(output), encoding="utf-8")
+    print(f"Wrote {len(repos)} rows for {current_year}")
 
-repositories_by_year = {}
-for year in range(current_year, current_year - 5, -1):
-    year_repositories = []
-    print(f"Fetching data for {year}")
-    response = requests.get(f'https://api.github.com/search/repositories?q=CVE-{year}%20in:name%20%20stars:>2%20language:Shell%20language:Go%20language:ASP%20language:WebAssembly%20language:R%20language:Lua%20language:Python%20%20%20language:C++%20language:C%20language:JavaScript%20language:Perl%20language:PowerShell%20language:Ruby%20language:Rust%20language:Java%20%20language:PHP&s=updated&o=desc&page=1&per_page=20')
-    if response.status_code != 200:
-        print(f"Failed to fetch data for year {year}: {response.status_code}")
-        continue
 
-    data = response.json()
-    total_count = data.get("total_count", 0)
-    print(f"Found: {total_count}")
-    total_repos_per_year[year] = total_count
-    if "items" in data:
-        items = data["items"]
-        if items:
-            year_repositories.extend(items)
-        else:
-            print(f"No more items found for year {year}")
-
-    if year_repositories:
-        # Sort the repositories by stargazers_count in descending order
-        #year_repositories.sort(key=lambda repo: repo['stargazers_count'], reverse=True)
-        repositories_by_year[year] = year_repositories
-
-# Define a class to handle repository information
-class RepositoryInfo:
-    def __init__(self, description, stargazers_count, name, html_url, updated_at):
-        self.description = description
-        self.stargazers_count = stargazers_count
-        self.name = name
-        self.html_url = html_url
-        self.updated_at = updated_at
-
-    def __hash__(self):
-        return hash(self.name + self.html_url)
-
-    def __eq__(self, other):
-        return self.html_url == other.html_url and self.name == other.name
-
-final_output = ['<h1 align="center">Recently updated Proof-of-Concepts</h1>']
-for year in range(current_year, current_year - 5, -1):
-    if year in repositories_by_year:
-        year_repositories = repositories_by_year[year]
-        year_repositories = [RepositoryInfo(repo["description"], repo["stargazers_count"], repo["name"], repo["html_url"], repo["updated_at"]) for repo in year_repositories]
-
-        final_output.append(f"\n\n## {year}\n")
-        final_output.append(f"### Latest 20 of {total_repos_per_year[year]} Repositories\n")
-        final_output.append("| Stars | Updated | Name | Description |")
-        final_output.append("| --- | --- | --- | --- |")
-
-        for repo in year_repositories:
-            try:
-                description = repo.description or ""
-                updated = time_ago(repo.updated_at)
-                final_output.append(f"| {repo.stargazers_count}⭐ | {updated} | [{repo.name}]({repo.html_url}) | {description} |")
-            except Exception as e:
-                print(f"Error generating final output for repository {repo.name}: {e}")
-                pass
-
-if repositories_by_year:
-    with open("README.md", "w", encoding="utf-8") as file:
-        file.write("\n".join(final_output))
-    print("Final output written to README.md")
+if __name__ == "__main__":
+    main()

.github/workflows/hot_cves.yml

@@ -24,6 +24,8 @@ jobs:
         cd /home/runner/work/cve/cve
         pip install requests
         python .github/getTrending.py
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         
     - name: Check for changes and commit if necessary
       run: |

.github/workflows/site.yml

@@ -12,6 +12,8 @@ on:
       - 'templates/**'
       - 'docs/assets/**'
       - 'README.md'
+      - '.github/getTrending.py'
+      - '.github/workflows/hot_cves.yml'
       - 'requirements.txt'
       - '.github/workflows/site.yml'