skills/garrytan/gbrain/schema-unify/Gen Agent Trust Hub

schema-unify

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use the gbrain CLI and jq for managing database schemas. These commands are used to check current versions, preview migration results, and submit processing jobs to a local or remote worker. These actions are appropriate for the skill's stated purpose of schema unification.\n- [INDIRECT_PROMPT_INJECTION]: The skill involves reading and acting upon data retrieved from tool outputs, creating a potential surface for indirect injection if database metadata were maliciously crafted.\n
  • Ingestion points: Data is ingested from gbrain onboard --check, gbrain schema stats, and gbrain jobs get as seen in SKILL.md.\n
  • Boundary markers: No specific prompt delimiters or "ignore" instructions are used to wrap tool outputs.\n
  • Capability inventory: The skill utilizes shell command execution, file system interaction via the CLI, and database modification capabilities.\n
  • Sanitization: The instructions do not specify explicit sanitization or validation of the data returned by the database tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 10:02 PM
Security Audit — agent-trust-hub — schema-unify