skills/jeremylongshore/claude-code-plugins-plus-skills/managing-database-recovery/Gen Agent Trust Hub
managing-database-recovery
Warn
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides multiple bash scripts for critical database management, such as failing over instances and performing point-in-time recovery. These scripts execute administrative commands with high privileges. \n- [COMMAND_EXECUTION]: The skill instructs the agent to establish persistence by creating and scheduling daily cron jobs to automate verification tasks. \n- [DATA_EXFILTRATION]: Instructions direct the agent to upload database backups to external S3 storage using the AWS CLI, facilitating the movement of potentially sensitive information to an external environment. \n- [CREDENTIALS_UNSAFE]: The skill requires the use of database superuser and cloud storage credentials. Automated handling of these secrets by an agent increases the risk of credential exposure. \n- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection by ingesting database configuration and status data. Ingestion points: database configuration files and system metadata outputs. Boundary markers: Absent. Capability inventory: File system access, network transmission via AWS S3, and shell script execution. Sanitization: No evidence of validation or escaping for the ingested system data.
Audit Metadata