Request Validation
Give Claude eyes on every request. AI validates inputs, detects injection attacks, and blocks malicious payloads at the edge.
Before
You
Are we properly validating user inputs? I'm worried about SQL injection.
I can review your code for validation patterns, but I can't see actual requests hitting your app.
Want me to audit your controllers for common vulnerabilities?
After
You
Are we properly validating user inputs? I'm worried about SQL injection.
Using nerve.get_blocked_requests
nerve.get_blocked_requests(type: "sql_injection", last: "24h")
Blocked Attacks (24h)
47 SQL injection attempts blocked
Most targeted: /api/users/search (23 attempts)
Pattern: UNION SELECT, OR 1=1
All blocked. Nerve caught 47 injection attempts at the edge. None reached your database.
Available via MCP (Model Context Protocol)
nerve.get_blocked_requests()
"What attacks were blocked?"
nerve.add_rule(pattern)
"Block requests with this pattern"
nerve.validate_schema(endpoint)
"Check this endpoint's validation"
nerve.get_suspicious()
"Show suspicious but allowed requests"
We're building Nerve as part of Phase 2. Follow along to get early access.