For fintechs
Ship fast. Pass due diligence.
Your engineers use AI agents. Your partner banks want proof they're controlled. Give them both.
- Free to start
- Runs offline
- Signed evidence
The questions banks ask.
And the answers you can give, with proof.
-
Do you know which AI agents and tools run on your laptops?
Yes. One read-only scan lists every agent, skill and MCP server.
-
Can an agent install anything it likes?
No. Known bad is blocked before it runs, by your policy.
-
Would you know if an approved tool changed?
Yes. Approved files are pinned by hash. A change is blocked.
-
Does your code go to another vendor?
No. Checks run on the laptop against a signed list.
-
Can you show us?
Yes. A signed evidence pack for SOC 2, ISO 42001 and OWASP.
Why they're asking.
- 341malicious skills in one campaignKoi, Feb 2026
- 1.5Magent tokens leaked by one skillWiz, 2026
- 437Kdownloads of one hijackable MCP bridgeCVE-2025-6514
- 0security fields in the MCP RegistryRegistry, 2026
Live in an afternoon.
-
Add the hook
One file for Claude Code or Cursor. See the setup
-
Set your policy
Block, ask a human or warn, per answer. Start warn-only.
-
Send the pack
Share the signed evidence pack with your bank's risk team.
Free to start.
-
Free
Builders
The list, the hook, the open API.
-
From $99/mo per team
Teams
One policy, one log, the evidence pack.
Planned pricing.
Good to know.
Is this a certification?
No. It is evidence of controls you run, signed so your partner bank can check it. Your auditor and your bank still make their own call.
Which agents does it cover?
Claude Code and Cursor through their pre-run hooks, and any pipeline through the command line.
What does it cost?
The list and the hook are free. Planned team pricing starts at $99 a month per team for one policy and one log across every agent.
Be ready before they ask.
Keep shipping with AI agents. Keep your bank partners close.