AI agents hunt every new threat inside your Sophos XDR, in SQL. Every match investigated, every empty result checked.
Looking for
Office app starts a script host
Running in Sophos XDR · SQL
SELECT meta_hostname, name, cmdline
FROM xdr_data
WHERE lower(parent_name) IN ('winword.exe', 'excel.exe')
AND lower(name) IN ('wscript.exe', 'mshta.exe')Example, simplified.
FourCore HUNT hunts every new threat that matters to you inside Sophos XDR, and brings back a verdict with its searches, evidence and limits.
01 · Hunt
A new campaign, an advisory or a plain question becomes searches written for Sophos XDR, run where your data already lives.
02 · Investigate
Follow-up searches run in Sophos XDR until each match is traced to normal software, admin activity, your own FourCore ATTACK tests, or nothing known.
03 · Check
Before "not seen", FourCore checks the logs, fields and retention behind each search. If Sophos XDR can't answer, the verdict says so and names what would close the gap.
Hunts run as SQL queries over your Sophos XDR Data Lake, so they search the data where it already lives.
Office app starts a script host
SELECT meta_hostname, name, cmdline
FROM xdr_data
WHERE lower(parent_name) IN ('winword.exe', 'excel.exe')
AND lower(name) IN ('wscript.exe', 'mshta.exe')SQL · example, simplified
Each relevant threat ends in a verdict, and every search behind it has a state.
With FourCore HUNT, AI agents turn a new threat, an advisory or a plain question into SQL searches, run them in Sophos XDR, follow every match to its cause and report a verdict with its evidence.
Give FourCore HUNT the report, advisory or indicators. It pulls out the behaviors and indicators, searches Sophos XDR for each one and tells you whether the threat was seen, with the searches behind the answer. Threats that matter to you can also be hunted on a schedule.
SQL. FourCore writes each search for Sophos XDR, so hunts run where your data already lives.
No. Hunts only read your data, and any action that would change something waits for approval.
The verdict says "could not check" for Sophos XDR and names the logging change that would close the gap. A blind spot is never reported as clean.
No. It writes the queries, checks the results and follows up on every match, and your hunters decide what happens next.
You pick the threats, we connect read-only, and we review the verdicts together.