cheat-migrate
Warn
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill's core logic involves reading natural language instructions from external files (e.g., migrations/registry.md and migrations/-to-.md) and executing them sequentially in Phase 3. This design creates an indirect prompt injection surface. 1. Ingestion points: migrations/registry.md and migrations/.md files (SKILL.md). 2. Boundary markers: No explicit delimiters or 'ignore embedded instructions' warnings are present to isolate migration content. 3. Capability inventory: The skill has access to Bash(), Write, and Edit tools, which are used to process migration steps. 4. Sanitization: No sanitization or validation of the natural language steps is performed before execution.
- [COMMAND_EXECUTION]: The skill utilizes the Bash(*) tool to perform file operations, such as creating backups and potentially executing commands specified in the migration steps. The combination of powerful shell access with instructions derived from external, potentially untrusted markdown files poses a risk of arbitrary command execution.
Audit Metadata