| name | royo-learn-onboarding |
| description | Initialize and verify royo-learn for a project, optionally install its integrations, then continue with capture-learning. |
| license | MIT |
| metadata | {"author":"RoyoTech","version":"2.0.0"} |
Royo Learn Onboarding
Use when
- royo-learn is being used in a project for the first time;
- an MCP call returns
project_not_found;
- the project marker or setup state is uncertain.
Workflow
- Identify the intended project root and check for
<root>/.royo-learn/config.yaml.
- If the marker is absent, run
royo-learn init --project-root <root> exactly once for that project root. Initialization is required; setup install is not a substitute.
- Run
royo-learn doctor --project-root <root> --json and resolve any reported failure before capturing a learning.
- Optionally run
royo-learn setup install --agent all --project-root <root> to register MCP and install skills for Claude Code, Codex, and OpenCode. This step is optional after initialization and is safe to repeat.
- Load and follow
../capture-learning/SKILL.md for capture eligibility, evidence, and output rules.
Clarifications
- Initialize one store per project root, not one per working directory. Project discovery walks upward from a subfolder until it finds
.royo-learn/config.yaml.
- Re-running init for the same root is idempotent; do not initialize every subfolder.
setup install configures supported agents and installs every valid directory under <root>/skills/; it does not create the project store.
- The royo-learn project store and Engram memory are separate, independent stores. Initializing, reading, or writing one does not initialize, synchronize, or replace the other.
Outcome
Return the project root, marker status, doctor result, whether optional setup ran, and whether control passed to capture-learning.