v1.0
L1 EssentialAutomated verification
3.4 URL and Resource Validation
3. Input Security — How do we protect against malicious input?
Description
URLs and external resources MUST be validated before the agent accesses them.
Rationale
Agents that fetch arbitrary URLs can be exploited for SSRF attacks.
Audit Procedure
1. Identify all code paths that fetch URLs 2. Check for protocol validation 3. Verify domain allowlisting 4. Test with internal IPs
Remediation
1. Implement URL allowlist 2. Block private IP ranges and cloud metadata 3. Validate protocols (https only) 4. Use URL parsing libraries
Framework Mappings
CIS Control 12NIST PR.AC-5