bear-migrate

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and transform untrusted data from files and databases, which creates a vulnerability where malicious instructions within the data could be processed by the agent.
  • Ingestion points: Data is read during the GATHER and MOVE phases from source code, configuration files, and database records across all domain guides.
  • Boundary markers: The instructions do not specify the use of delimiters or clear directives to the agent to ignore instructions found within the data being migrated.
  • Capability inventory: The agent utilizes powerful tools including git, grep, sqlite3, wrangler, and pg_dump to perform its tasks.
  • Sanitization: The migration patterns do not include explicit sanitization or validation of the content being transformed.
  • [DYNAMIC_EXECUTION]: The skill involves generating and executing dynamic code to perform migrations, specifically database migration scripts and file transformation logic.
  • Script generation: The domain-database.md guide provides templates for creating Kysely migration scripts and batch processing functions at runtime.
  • Execution: The agent is instructed to run these generated scripts and shell-based transformation commands (e.g., sed loops) to execute the migration.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 01:31 AM