database-backups

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides templates for shell scripts that interact with database management systems and cloud storage. It references standard tools including pg_dump, mysqldump, pg_restore, psql, docker, and the aws CLI. These operations are consistent with the skill's primary purpose of implementing database backup strategies.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a vulnerability surface by providing script templates that an agent might populate with user-supplied data, such as database names or backup locations.
  • Ingestion points: Parameters within the shell script templates in SKILL.md (e.g., database names, environment variables).
  • Boundary markers: No delimiters or specific safety instructions are present to prevent command injection into the generated scripts.
  • Capability inventory: The skill utilizes subprocess calls for database operations, container management via Docker, and network uploads via the AWS CLI.
  • Sanitization: No sanitization or validation logic is provided within the template examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 02:05 PM