Skill Instructions

Inputs / Outputs / Non-goals

• Inputs: JSON schemas, affected feature specs, validation commands.
• Outputs: schema-safe JSON edits, impact notes, suggested validations.
• Non-goals: schema-breaking changes without explicit approval.

Trigger conditions

Use this skill when prompts include or imply:

• Editing judoka.json, tooltips.json, or battle/config JSON.
• Adding metadata/config fields consumed by runtime code.
• Updating JSON-backed content contracts.

Mandatory rules

• Preserve schema stability and prefer additive changes.
• Provide explicit defaults for new fields.
• Identify impacted consumers before finalizing edits.
• Treat JSON as executable configuration and avoid ambiguous keys.

Validation checklist

• Run npm run validate:data after JSON edits.
• Run core checks: npm run check:jsdoc && npx prettier . --check && npx eslint ..
• Run targeted tests for impacted consumers when behavior changes.

Expected output format

• JSON change summary with rationale.
• Consumer impact analysis and required follow-up code updates.
• Validation command results.

Failure/stop conditions

• Stop when requested JSON edits are schema-breaking without approval.
• Stop when downstream impacts cannot be determined confidently.