Aabyss-Team/Threat-Broadcast-Plus
Security & PrivacyAggregates public vulnerability feeds into a static threat-intelligence page with optional AI risk briefs.
What it does
Threat-Broadcast-Plus is a Python crawler and static-site generator for security vulnerability intelligence. Every run pulls recent entries from several public sources, de-duplicates them, stores them in SQLite, and renders a searchable static page under docs/index.html. A GitHub Actions workflow can run it every four hours and commit the regenerated page and database back to the repository.
Enabled sources include Alibaba Cloud AVD high-risk list, Tophant intelligence, ThreatBook, Chaitin VulDB, NVD's 2.0 recent JSON feed, Tenable CVE RSS, and the CISA KEV catalog README source. Each source is fetched independently, so one failure does not stop the rest. Time fields are normalized to UTC+8, and duplicated entries are keyed by source, identifier, and link.
If a Kimi API key is configured, the script sends a filtered set of recent entries to the model and writes a short Chinese risk brief to data/ai_summary.json, which appears in a card on the page README source. Without a key, the page still generates and the previous summary is preserved.
Quick start
Requirements: Python 3.8 or newer (CI uses 3.12). From the repository root:
python -m pip install -r requirements.txt
python main.py -g AabyssTeam-Plus-F5m5bAxpaHE2CwS -t 60 --no-mail --no-wechat
The generated page is docs/index.html. Because the page uses a strict CSP, preview it through a local static server rather than opening the file directly:
python -m http.server 8000
# open http://127.0.0.1:8000/docs/
The -g value is the startup password; local runs must match the main.py default, while the GitHub Actions workflow expects the CRAWL_PWD secret README source. The workflow also accepts optional secrets for NVD, mail, WeChat push channels, and Kimi.
For unattended operation, enable the included autorun.yml workflow and set GitHub Pages to deploy from the default branch's /docs directory. A custom domain is read from docs/CNAME.
A ready-to-browse deployment is available at the online broadcast page.
Configuration and retention
Main settings live in config/settings.yml: database path, retain_days (default 365, 0 disables time-based cleanup), max_rows_per_src (default 3000), per-source fetch limit top, HTTP timeout, SMTP and WeChat notification credentials, and Kimi model parameters README source.
The page shows only the latest top entries per source (default 60) and hides sources that have not updated for more than 45 days. The SQLite database is written to data/cves.db; cache/*.dat files hold MD5 values for cross-run de-duplication and can be deleted safely because the database unique index also prevents duplicates.
Constraints
- Only CVEs already analyzed and published by the listed vendors are collected; this is not a complete CVE mirror. The README points readers to MITRE, CVEProject, and CIRCL for full coverage README source.
- Some source detail pages require login, notably ThreatBook; the list endpoints themselves are public.
- Kimi keys must match their endpoint: Kimi Code console keys use
https://api.kimi.com/coding/v1/chat/completionswith modelk3, while Moonshot platform keys use theapi.moonshot.cnorapi.moonshot.aiendpoints README source. - The default workflow commits
data/cves.dbanddocs/index.htmlevery four hours, so the repository grows over time; the README suggests moving the database to Actions cache and deploying the page as a Pages artifact if that matters. - The project is GPL-3.0 and is derived from earlier threat-broadcast forks README source.
HotRepo's take
The repository has 108 stars and 25 forks, with an early new-project snapshot recording 6 stars and 2 forks gained in a one-hour window on 2026-11-19; no GitHub daily rank has been recorded. That is a small sample, so it does not establish a trend. What stands out from the material is the maintenance-oriented rework: standard-library SQLite and logging replace abandoned dependencies, NVD moved to the official 2.0 JSON feed, several Chinese vendor sources were added, and GitHub Actions was refreshed for Python 3.12. Those are practical fixes for a crawler that has to keep working as upstream sites change.
Written by AI from the project's README, updated . The project's own docs are the source of truth.
You might also like
- OpenCTI-Platform/openctiA self-hosted platform to structure, link, visualize and share cyber threat intelligence using STIX2.
- NVIDIA/SkillSpectorChecks an AI agent skill for malicious patterns and vulnerabilities before you let an agent run it.
- Haryy985/yunkun-console-previewA preview console that keeps authorized pentest tasks, agent traces, HTTP evidence and findings in one workspace.