rye-installer

name	rye-installer
description	Install, migrate, verify, and test Rye PostgreSQL deployments. Use when a user asks to bootstrap Rye in a database, apply core/profile migrations, run schema verification, seed sample data, or execute conformance checks.

Rye Installer

Workflow

1. For a new local trial, run the public bootstrap. Prefer https://projectrye.dev/onboard; if that hosted endpoint is unavailable, fall back to the GitHub-hosted script:

   tmp="$(mktemp)"
   curl -fsSL https://projectrye.dev/onboard -o "$tmp" || \
     curl -fsSL https://raw.githubusercontent.com/BLTGV/project-rye/main/site/public/onboard -o "$tmp"
   sh "$tmp"
   rm -f "$tmp"
   

2. For a remote PostgreSQL database, run the same bootstrap with --remote "$DATABASE_URL":

   tmp="$(mktemp)"
   curl -fsSL https://projectrye.dev/onboard -o "$tmp" || \
     curl -fsSL https://raw.githubusercontent.com/BLTGV/project-rye/main/site/public/onboard -o "$tmp"
   sh "$tmp" --remote "$DATABASE_URL"
   rm -f "$tmp"
   

3. Install the agent onboarding skill with npx skills add BLTGV/project-rye --skill rye-onboarding.

4. Start the first adoption unit with ./scripts/rye onboard --label ... --purpose ....

5. Use ./scripts/rye status to confirm the selected database and active onboarding scope.

6. Use the lower-level install/conformance commands only when debugging migrations or release gates.

The first-run path writes .rye.env so later commands use the same database without the user re-entering the connection string.

Install always applies schema migrations and syncs plugin metadata from plugins/*/rye-plugin.json. That metadata includes contributed node, edge, assertion, event, and artifact types plus onboarding questions, never-infer defaults, validation hooks, and admin hooks. --seed is only for quickstart example data.

Commands

• Local trial: curl -fsSL https://projectrye.dev/onboard | sh
• Local fallback: curl -fsSL https://raw.githubusercontent.com/BLTGV/project-rye/main/site/public/onboard | sh
• Remote install: curl -fsSL https://projectrye.dev/onboard | sh -s -- --remote "$DATABASE_URL"
• Remote fallback: curl -fsSL https://raw.githubusercontent.com/BLTGV/project-rye/main/site/public/onboard | sh -s -- --remote "$DATABASE_URL"
• Agent skill: npx skills add BLTGV/project-rye --skill rye-onboarding
• Repo-local trial: ./scripts/rye local --fresh
• Repo-local remote install: ./scripts/rye remote --db-url "$DATABASE_URL"
• Plugin metadata: ./scripts/rye plugins list
• First scope: ./scripts/rye onboard --label "First Scope" --purpose "..."
• Status: ./scripts/rye status
• Core only: ./scripts/install.sh --profiles ''
• Core + CRM: ./scripts/install.sh --profiles crm
• Core + PM: ./scripts/install.sh --profiles pm
• Core + CRM + PM: ./scripts/install.sh --profiles crm,pm
• Full checks: ./scripts/conformance.sh

Post-Install

After install, run SELECT rye_catalog() to confirm the instance is ready and see what's connected.

Supabase Deployment

Rye can be deployed to Supabase via its MCP apply_migration or execute_sql tools:

1. Create missing extensions: btree_gin, pg_trgm (pgcrypto is pre-installed).
2. Apply each migration file (0001 through 0007, plus profiles) via apply_migration.
3. Verify with SELECT rye.rye_catalog().

Key differences from self-hosted:

• postgres is not a superuser — RLS applies to all queries. Always set session vars.
• Session variables reset per call. Prefix every query with set_config() calls.
• SET app.current_role = 'admin' syntax fails through the MCP. Use set_config('app.current_role', 'admin', false) instead.

Notes

• Target PostgreSQL 15+.
• Rye installs alongside existing tables. It does not modify them.
• Use link_record() to connect existing domain table rows to the graph.
• Use track_table() to attach CDC triggers for change tracking.
• Assertion updates are function-only; verify this via ./scripts/verify.sh and security tests.
• Do not edit migration order in schema/migrations without versioning and tests.
• If migrations fail partway, fix the root cause and re-run ./scripts/migrate.sh.