<< All versions
Skill v1.0.1
currentAutomated scan100/100redhatproductsecurity/prodsec-skills/authentication
+2 new
──Details
PublishedAugust 1, 2026 at 06:06 PM
Content Hashsha256:9aec98fbdcb0066f...
Git SHAb55c6b9c3392
Bump Typepatch
──Files
Files (1 file, 1.7 KB)
SKILL.md1.7 KBactive
SKILL.md · 42 lines · 1.7 KB
version: "1.0.1" name: authentication description: Enforce standard authentication for external data source connections. Use when integrating, configuring, or reviewing connections from AI systems to external databases, APIs, or data services. metadata: category: secure_development subcategory: external-data-source
Authentication for External Data Sources
Security Requirement
External data sources MUST require that the principal (user or service) connecting to them is identified and authenticated. The authentication MUST use standard protocols.
Standard Authentication Protocols
| Protocol | Use Case | |
|---|---|---|
| OAuth 2.1 / OIDC | Web APIs, REST services, cloud data sources | |
| SPIFFE/SPIRE + mTLS | Service-to-service connections within infrastructure | |
| SAML | Federated enterprise data sources | |
| Kerberos | On-premise databases and services (Active Directory environments) | |
| Client certificates (mTLS) | Machine-to-machine data source access |
What Is Not Acceptable
- Unauthenticated connections to data sources containing any data
- Shared credentials used by all services
- Credentials embedded in connection strings in source code
- Long-lived API keys without rotation (see
api_keys/avoid-api-keys)
Implementation Checklist
- [ ] Identify all external data sources accessed by the AI system
- [ ] Require authentication for every data source connection
- [ ] Use standard authentication protocols (OAuth, mTLS, Kerberos)
- [ ] Use unique credentials per service or workload (no shared credentials)
- [ ] Store credentials in a secret management system
- [ ] Rotate credentials on a regular schedule
- [ ] Log authentication events (successes and failures)