One command. Every agent under your rules.
Claude Code, Cursor, Windsurf, Copilot and Codex follow your rules from their next session. No pipeline to build.
npx -y @cybedefend/vibedefend@latest install
Claude CodeWired
CursorWired
OpenAI CodexWired
GitHub CopilotWired
WindsurfWired
Your rules, applied at the line.
Business and security rules arrive as the agent writes: 89% applied exactly, 13% with a rules file.
Every query is scoped to the tenant.
rules/data-access.mdawait db.invoices.find({ id, tenant: req.tenant })Scoped to the tenant, at the line
Dangerous actions never run.
DROP TABLE, sudo, a .env read: refused before it starts, and logged for your team.
Cursor wants to runDROP TABLE invoices;
- Read
- Checked against your guards
- Refused
- Run
Stopped before it ran.
Fixed before the commit.
When the agent finishes, its changes are scanned and it fixes its own findings. Your reviewers read clean code.
SQL injection, reachable
CWE-89, src/db/users.repo.ts:42"SELECT * FROM users WHERE id = $1", [id]Fixed by the agent before the commit
Then proved, not assumed.
After the commit, BLSA checks its business logic and proves each verdict: the rule, the path, the line.
Refunds above €500 need a second approver.BLSA, proof mode
POST /refundsEntry point
approvals.second(order)The rule's check
payments.refund(order)Runs after the check
Holds: approved before it is paid
Intelligence is solved. Context isn't.One command brings your business logic and security into your
npx -y @cybedefend/vibedefend@latest install✓Fixed in the same session
Less rework. Measured.
One agent, thirty tickets, three setups, blind audits. Rerun it yourself.
8×
fewer broken rules per ticket 2.28 with no tool, 2.27 with a rules file, 0.28 with VibeDefend
~$400,000
saved a year, for twenty developers computed from the measured rates
5
coding agents wired by one command Claude Code, Cursor, Windsurf, Copilot, Codex
Everything your agents do, on one map.
PaymentsPlatformMobileDataGrowthCorecheckout-apiledgerbilling-webpayouts-workerfraud-rulesapi-gatewayauth-serviceadmin-consoleinfra-terraformios-appandroid-appmobile-bffevents-pipelinewarehouseml-featuresreportingweb-appexperimentsmarketing-siteuser-servicenotificationssearch-api
Day 1: rules, guards and scans, live everywhere.
Day 30: every rule served and every action blocked, ready for your audit.
A demo organisation: its names, people and figures are fictitious.They ship with it every day.
Vulnerability analysis and remediation have become significantly faster. We save valuable time every day.
OlivierIt secures our projects by detecting and fixing vulnerabilities in code and dependencies, with seamless integration.
GeoffreyWe tried the same approach with other AIs and never got results this convincing.
Alexandre Lignon
Every agent. One context.
One command wires every agent on your machine to the same rules, guards and checks.
npx -y @cybedefend/vibedefend@latest install- Claude Code
- Cursor
- Windsurf
- GitHub Copilot
- OpenAI Codex
In detail
How it works, step by step
Wired.
Served.
Guarded.
Checked.
Proved.
Learned.
One developer's day, as the team sees it
Per developer, not per repository: the rules each agent was served, and what it was stopped from doing.
- Served 3 rules on
src/api/refunds.ts - Tried
rm -rf ./migrations - Tried to read
.env, used the managed reference - Diff checked before the commit, one finding fixed
- Proposed a rule: refunds above €500 need a second approver
Why a rules file is not enough
A rules file does not fix it. The rule was in the file, word for word, and the agent broke it anyway.
2.27
rules broken per ticket with a rules file kept by hand. Without any file: 2.28.
Controlled study, August 2026
What people ask first.
What is CybeDefend?
CybeDefend gives AI coding agents the context they lack. Your business rules and security rules reach the agent at the line it edits, destructive commands and secret reads are refused before they run, and every change is scanned before the commit, with its fix. VibeDefend, its agent layer, works inside Claude Code, Cursor, Windsurf, GitHub Copilot and OpenAI Codex.
How is it different from a code scanner?
A scanner reads code after it is written and sends alerts to people. CybeDefend works while the agent writes: it hands the agent your rules before the edit, stops what must never run, and gives the agent one finding it can fix in the same session. The scan is still there, on the change, before the commit.
How do I install it?
Run one command in any terminal, on macOS, Linux or Windows, with Node 18.17 or later. The installer finds the coding agents on your machine and wires each of them; you choose your region, EU or US, and sign in once in your browser.
Does it work? What was measured?
In CybeDefend's controlled study of 24 August 2026, one agent worked the same thirty tickets in three setups. With VibeDefend it applied 89% of the precise rules, against 12% with no tool and 13% with a rules file kept by hand. The protocol, the transcripts and the blind audits are public.
My agent already checks its own code. Why add CybeDefend?
Each agent checks its own work, with its own settings. CybeDefend gives all five agents the same rules, your business rules included, served before the edit rather than reviewed after it, and keeps one record of what each agent was served, what was blocked and what was fixed.