debug prod issues
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: Accesses production environment variables and database records.
- Loads sensitive configuration from '.env.production.local'.
- Connects to production Neon Postgres databases to inspect user and workspace data.
- [COMMAND_EXECUTION]: Executes various CLI tools and dynamic scripts for environment management.
- Uses 'vercel-cli' for log streaming, environment variable management, and production deployments.
- Executes TypeScript code dynamically using 'pnpm tsx -e' for database inspection tasks.
- Uses 'git' for repository synchronization and 'curl' for API endpoint testing.
- [PROMPT_INJECTION]: Vulnerability surface when processing untrusted production data.
- Ingestion points: Production logs from 'vercel logs' and database query results.
- Boundary markers: None identified in the provided file; the agent is not instructed to isolate or ignore commands within logs.
- Capability inventory: File system access (reading/writing .env files), shell command execution, and database connectivity.
- Sanitization: No validation or sanitization of data retrieved from external logs or database fields is defined.
Audit Metadata