research.assist()
Look deeper.
We use language models to review code, triage potential issues, and draft clear reports.
RESEARCH WITH PERMISSION.NO AUTOPILOT. NO SHORTCUTS.
researcher@mny:~$ whoami
Question
everything.
Stay in scope.
Independent security researchers & bug bounty hunters. We find vulnerabilities in authorized programs and report them so they get fixed.
hello@mny.lolMODEL-ASSISTED. HUMAN-VERIFIED.
01 / METHODcat approach.md
Tools do the heavy lifting.
People make the call.
Language models help us look closer.
Responsibility stays with us.
human.verify()
Prove the finding.
A person checks every finding before submission. Models are tools, never autopilot.
report.private()
Help get it fixed.
Actionable reports go privately to the affected program, with time for the vendor to fix.
02 / RULES OF ENGAGEMENTcat boundaries.conf
// NON-NEGOTIABLE
Permission first.
Always.
Good research respects the system
and the people behind it.
Explicit scope.
Only in-scope targets, under each program’s published rules.
Respect the limits.
Follow rate limits and testing restrictions. Stay within the agreed boundaries.
Minimum data.
No access to or retention of data beyond what’s needed to demonstrate an issue.
Coordinated disclosure.
Report privately. Give the vendor time to fix before anything is public.
03 / CONTACTstart a conversation
researcher@mny:~$ mail hello@mny.lol
Let’s compare notes_
hello@mny.lolFor programs, questions, and responsible disclosure.