| name | security-review |
| description | Security analysis for Azure-hosted Python services. Checks for managed identity, secret exposure, input validation, RBAC, and OWASP-style vulnerabilities. Use when reviewing endpoints, auth code, infra changes, or any user-input handling. |
Security Review
Perform a security-focused review adapted to the KB Agent project — an Azure-hosted pipeline with managed identity everywhere.
Risk Context
This project processes knowledge base content and serves it via a conversational agent. Security issues can lead to:
- Unauthorized access to Azure resources (AI Search, Storage, Cosmos DB)
- Data leakage from the knowledge base
- Prompt injection via user queries to the agent
- Path traversal via blob storage operations
Checklist
Authentication & Authorization
Input Validation
Secrets & Configuration
Infrastructure Security
Dependencies
Output Format
Rate each finding:
- CRITICAL — Exploitable vulnerability, must fix immediately
- HIGH — Significant risk, fix before merge
- MEDIUM — Defense-in-depth issue, should address
- LOW — Best practice suggestion
Include specific file + line references, the vulnerability type, and a concrete fix.