Detects the gbrain engine and migrates PGLite brains to Postgres or Supabase.
npx skills add https://github.com/garrytan/gbrain --skill postgres-adopt
Please help me install this Agent Skill. Skill: postgres-adopt Source: https://github.com/garrytan/gbrain/tree/main/plugin/skills/postgres-adopt Command: npx skills add https://github.com/garrytan/gbrain --skill postgres-adopt
Operators running gbrain often don't know whether their brain is backed by PGLite or Postgres, and switching engines by hand risks splitting data across two stores. This Skill detects the active engine, provisions Postgres when wanted, and migrates existing PGLite data through the guarded migration path.
gbrain engine status --json to report the effective engine, config source, and probe health without needing the database up.gbrain init --prefer-postgres to walk an install ladder covering env URLs, Supabase Management API discovery, local Postgres, opt-in Docker, and PGLite fallback.gbrain migrate --to supabase, which resumes partial runs from a manifest and only flips config after a fully clean run.gbrain doctor.Ask the agent to check which gbrain engine is running and upgrade the brain to Postgres if it is still on PGLite.
High-intent search queries and answers about installing and using this skill.
Run `gbrain engine status --json` to see the effective engine, its config source, and probe health. The command works even when the database is down, which is the point of the detection step.
Use `gbrain migrate --to supabase --url <postgres-connection-string>` after confirming with the operator. A partial run leaves you on PGLite and resumes from its manifest when re-run; config only flips after a fully clean run.
Postgres wins on concurrency, multi-machine access, and brains over 1000 pages. PGLite keeps the per-turn bootstrap hook lane, which is PGLite-only, so it remains a fine choice for single-machine setups.
No. `gbrain config set engine` is refused by design because flipping the engine without migrating data splits the brain across two stores. Always use the migrate command, which brings its own guardrails.
Yes. Set SUPABASE_ACCESS_TOKEN for Management API discovery, plus SUPABASE_PROJECT_REF on multi-project accounts and SUPABASE_DB_PASSWORD for the connection string. The init ladder tries this rung before local Postgres and Docker.
An environment URL is overriding the config file. Check the `db_url_source` field: `env:GBRAIN_DATABASE_URL` or `env:DATABASE_URL` means the env value wins over the file setting.
SakshamKandelConfigure gbrain with PGLite or Supabase and register MCP.
yunusgungorGenerate a JSON health report combining gbrain doctor and migration status.
iota-uzConnect to local or staging PostgreSQL databases and run diagnostics.
zhu637882-stackInstall the gbrain CLI, initialize a PGLite or Supabase brain, and register an MCP with per-remote trust policy.