validate-schema

Validate SpiceDB Schema

Validate any SpiceDB .zed schema file using the schema-validator agent. Works on any .zed file -- plugin-generated, manually edited, or migrated from another system.

Process

Step 1: Locate Schema

If the user provided a schema file path, use that. Otherwise:

1. Search for *.zed files in the current working directory
2. If exactly one is found, use it
3. If multiple are found, ask the user which to validate using AskUserQuestion
4. If none are found, ask the user for the schema file location

Step 2: Launch schema-validator Agent

Use the Task tool to launch the schema-validator agent:

Task(
    subagent_type="spicedb-dev:schema-validator",
    description="Validate SpiceDB schema",
    prompt="Validate the schema file at [path] and suggest improvements"
)

The agent will:

• Check syntax for .zed schema language validity
• Validate best practices (naming conventions, relation/permission structure)
• Suggest optimizations
• Identify potential issues (anti-patterns, missing type annotations, etc.)

Step 3: Present Results

Present the validation results to the user with:

1. Summary: Pass/fail and number of issues found
2. Issues: List each issue with its severity (error, warning, suggestion) and location
3. Next steps:
   • If issues found: describe what to fix and how
   • If no issues: suggest running /spicedb-dev:generate-schema (if no schema yet) or /spicedb-dev:implement-spicedb-checks to add authorization code

Tips

• This command works on any .zed file, not just plugin-generated schemas
• Run after manually editing a schema to catch regressions
• Run after migrating a schema from another system to verify correctness
• The generate-schema command runs this automatically after generation

Notes

• The schema-validator agent performs static analysis only -- it does not connect to a live SpiceDB instance
• For runtime validation with actual relationships, use spicedb serve-testing (see authorization-testing skill)