bdc-validate-share-metadata

You are running the /bdc-validate-share-metadata command.

The user wants to validate ORD (Open Resource Discovery) metadata before pushing it through create_or_update_share. The ORD payload to validate is in $ARGUMENTS — it may be either:

• An inline JSON object, e.g. {"title": "...", ...} or {"@openResourceDiscoveryV1": {...}}
• A path to a .json file on disk

If $ARGUMENTS is empty, ask the user to paste the ORD JSON or provide a file path, then stop.

Steps

1. Resolve $ARGUMENTS into a Python dict — read from disk if it looks like a path, otherwise json.loads() it.
2. Call the validate_ord_metadata tool from the sap-bdc MCP server with the parsed object as ord.
3. Present the result in this exact shape:

### ORD metadata validation

Status: ✅ Valid  /  ❌ Invalid

Errors (block publish):
  - <each error>

Warnings (non-blocking):
  - <each warning>

Spec reference: https://sap.github.io/csn-interop-specification/

Source rules: SAP help.sap.com "Working with Data Products in SAP Databricks" (May 2026), pages 13–15.

1. If valid=true, suggest the next step: call create_or_update_share (which in v0.5.0+ will re-run this validator automatically; pass skip_validation=true only if you want to bypass).

2. If valid=false, do NOT call create_or_update_share. List the rules the ORD violates and propose a fix:

   • Required field missing → ask the user to supply it
   • description contains shortDescription → suggest a rewrite that doesn't quote shortDescription verbatim
   • Invalid enum (visibility, releaseStatus) → list the allowed values
   • Bad ISO 8601 date → show the expected format (2026-12-19T15:47:04+00:00)
   • sunsetDate < deprecationDate → call out which one to fix

Do NOT modify the user's ORD without confirmation. The goal is to surface validation failures before a round-trip to the SDK.