create-migration
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements comprehensive security guardrails (ASI03-ASI09) that explicitly address common attack vectors such as identity abuse, remote code execution, and context poisoning. It mandates safe database practices (lock timeouts, concurrent indexing, batching) and requires validation of data before use.
- [INDIRECT_PROMPT_INJECTION]: The skill processes local files and database content, creating a surface for indirect instructions. However, it includes mitigations such as SQL linting (ASI05) and context validation (ASI06). Ingestion points: reads repository files and database schema/data using read_file, search_code, and query_db. Boundary markers: none explicitly defined for data ingestion. Capability inventory: file system modification and database query execution. Sanitization: linting migration SQL against patterns and validating notes against the live schema.
Audit Metadata