database-administrator

Warn

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes instructions and scripts for high-privilege system modifications, specifically using sudo for installing database software and configuring system services in REFERENCE.md.
  • [COMMAND_EXECUTION]: Implementation of persistence mechanisms through the configuration of crontab for automated database backup schedules as described in the deployment steps of REFERENCE.md.
  • [COMMAND_EXECUTION]: Shell scripts for backup verification in EXAMPLES.md perform directory management operations including tar extraction and recursive removal (rm -rf) on temporary paths.
  • [EXTERNAL_DOWNLOADS]: Fetches system dependencies and management utilities using standard package managers (apt, pip) and performs data synchronization with AWS S3 for continuous WAL archiving and base backups in REFERENCE.md.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection when analyzing database statistics and logs to troubleshoot or optimize performance. Ingestion points: Untrusted data enters the context from database system tables such as pg_stat_statements and pg_stat_user_tables in EXAMPLES.md. Boundary markers: Absent; there are no delimiters separating data retrieved from the database from the agent's instructions. Capability inventory: Extensive capabilities including system-level execution via sudo, crontab modification, and network operations via AWS CLI. Sanitization: No explicit validation or sanitization is implemented for data retrieved from external database sources.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 21, 2026, 08:31 AM
Security Audit — agent-trust-hub — database-administrator