Detecting critical CVEs without a SIEM: a practical Sigma rules workflow for small teams
When a critical CVE drops — say a CVSS 9.8 in something you actually run — the window between disclosure and exploitation is measured in hours, not days. Enterprise teams have detection engineers and SIEM pipelines for t
When a critical CVE drops — say a CVSS 9.8 in something you actually run — the window between disclosure and exploitation is measured in hours, not days. Enterprise teams have detection engineers and SIEM pipelines for this. A two-person team or a solo freelancer has… a log file and anxiety.
I got tired of watching critical vulnerabilities trend on Hacker News while the practical advice was always "deploy a SIEM." So I built a workflow that runs Sigma rules against plain logs on a laptop — no infrastructure, no license, no team.
The core idea
Sigma is a generic, YAML-based detection format. The same rule works across log sources — you just need a converter (sigmac/pySigma) to translate it into whatever you're actually querying: SQLite, Elasticsearch, Splunk, whatever.
The workflow:
- Collect — ship the logs you already have (auth logs, web server logs, app logs) into SQLite or flat files
- Convert — pull Sigma rules for the CVE/class you're worried about, convert to your backend
- Hunt — run the converted queries on a schedule or on-demand
- Triage — hits go through a severity checklist before anyone gets paged
For most small environments this runs in under a minute and catches the things that matter: webshells, suspicious auth patterns, exploitation attempts against the specific CVE.
What actually works
After running this against real log data, a few things surprised me:
Rule quality beats rule quantity. The SigmaHQ repo has thousands of rules, but for a given CVE maybe 5–10 are actually relevant to your stack. Curating beats importing everything — noisy rules train you to ignore alerts, which is worse than no alerts.
The CVE news cycle is your friend. Critical CVEs generate detection rules within hours because the community moves fast. Subscribing to a curated feed (or just watching HN's front page for security tags) tells you exactly when to pull fresh rules.
Triage is where small teams win. Enterprise teams drown in alerts and automate triage badly. A two-person team with a 5-question checklist (Is this asset internet-facing? Is the rule's confidence high? Does the log source even exist here?) makes faster decisions.
Where I'd start
If you want to try this today:
-
pySigma is the current standard converter —
pip install pysigmaand you're converting rules in minutes -
SigmaHQ on GitHub is the canonical rule repo — filter by
level: criticaland your tech stack - Start with web/auth logs only. Don't try to cover everything on day one.
I packaged the whole workflow — curated starter rules for the CVEs that actually matter, the conversion commands, and the triage checklist — into a Sigma Rules Detection Pack. It's $23, it's a zip file, and it takes about 20 minutes to get running. Pay in BTC if you want, no account needed:
→ https://choosing-meet-subscribers-rankings.trycloudflare.com/?sku=sigma-rules-pack
If you'd rather start free, the agent security checklist has the log-collection basics: https://choosing-meet-subscribers-rankings.trycloudflare.com/free
What's the messiest log situation you've had to debug at 2am?
Originally published by Dev.to Security. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.