CAIN-42 CAIN StudioAGENTIC AI 75 · #23
Autonomous vulnerability investigation and repair
Google described Project Fortify workflows spanning security scanning, triage and remediation.
Last reviewed 2026-11-11
Agentic AI 75 / #23
2026 status: Vendor-reported deployment · Category: Security engineering
The 2026 development
Google described Project Fortify workflows spanning security scanning, triage and remediation.
The governance gap
An agent that can find and patch vulnerabilities can also patch the production path without review. Speed is useful; unreviewed write access to a security control is not.
Where CAIN-42 fits: Moat 6 — Adversarial Trust Engine
Core question: Can CAIN break its own guarantees? Continuous attack, mutation and red-team testing of the controls themselves.
Keep detection deterministic, make the fix a proposal, and prove it with a test the fixer did not author. CAIN's MCP scanner and KEYSTONE loop already separate the detector, the actor and the independent prover.
How the wiring works
- Run the deterministic detector and record its verdict and rule version.
- Have the agent propose a minimal patch with its rationale.
- Run an independent reference test on the original and the patched tree.
- Sign the result only when every check holds; otherwise label it not verified.
What is real today
LIVE A public route in this deployment exercises this mechanism today, within its stated scope.
The hosted scanner is heuristic and the hardened fixes are bounded samples. It is not Project Fortify and does not scan a customer production estate.
| Component in this repository | What the module says it is |
|---|---|
mcp-security-scanner/engine.py | Real MCP tool-schema security scanner. |
keystone/mcp.py | MCPGate KEYSTONE: the MCP security scanner finds a poisoned or unsafe MCP tool definition, Claude hardens it, |
keystone/redteam.py | REDTEAM KEYSTONE: attack an agent's tool policy, write the smallest patch, prove the patch closes every breach. |
keystone/engine.py | KEYSTONE: FALSEWORK finds what is fake, Claude fixes it, and every fix is proved and signed. |
Descriptions are the first line of each module's own docstring, read without importing it. A module path is a pointer into this repository, not a hosted endpoint. Its existence is checked by the test suite; its status for this use is the label above.
Check it yourself
- Scan a live MCP tool definition
/mcpgate-keystone - Fire a live decision through the CAIN-42 pipeline
/fabric/try— the same enforcement path every product sits behind.
Commercial package
| Offer | Vulnerability-to-Patch Proof |
|---|---|
| Buyer | Security engineering and AppSec teams |
| Price (proposal) | $2,500–$5,000 per engagement |
| Bundle | Code & Proof — Prove the fix, not just the diff. |
| Operator stream | Stream 5 of the 10-stream operator playbook |
Platform engineering, AI coding-tool vendors, security engineering is the bundle buyer. The entry offer for this bundle is Fabrication audit or long-run engineering attestation; the range is $1,500–$6,000 per audit; $2,500–$4,000/month retainer. A price on this page is a proposal, not a validated rate, and no customer is claimed.
Buy this as a product or service
CAIN-A75-023 — Vulnerability-to-Patch Proof DELIVERABLE
A bounded engine run on the buyer's sample: findings, a written report, and a SHA-256 artifact manifest. The manifest is unsigned: it is not a signature and not an attestation.
Kind: Fix + Prove · Price (proposal): $2,500–$5,000 per engagement · Lead time: 5 business days.
Open the product page · all 75 products
What this page does not claim
What this page is not. CAIN-42 is not affiliated with, endorsed by, integrated into, or shipped with the organisation named in the 2026 item. The item is public context; the page describes how CAIN-42 would govern that class of capability. No page here claims a customer, a partnership, a certification, a patent, or a production integration. Prices are proposals. Built by a three-person team.