db-migrate
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill relies on external configuration files (
docs/project-config.jsonandbackend-patterns-reference.md) to define migration commands. If these files are maliciously modified in a repository, the agent would execute those commands during the migration workflow. - Ingestion points: The skill reads from
docs/project-config.json,backend-patterns-reference.md, and processes user input via$ARGUMENTSinSKILL.md. - Boundary markers: There are no explicit boundary markers or warnings to ignore instructions within the project configuration files or the user-provided arguments.
- Capability inventory: The skill uses shell execution capabilities to run migration tools (
bashblocks inSKILL.md). - Sanitization: No evidence of sanitization or validation of the configuration values before they are interpolated into the shell command placeholders (e.g.,
{configured-migration-add-command}). - [COMMAND_EXECUTION]: The skill is designed to construct and execute shell commands for database operations. It relies on the environment to provide the necessary migration tools, which grants the agent the capability to interact with the system's CLI based on local project settings.
Audit Metadata