sc4sap:sap-doctor

SC4SAP Doctor

Diagnoses the full SC4SAP stack: plugin installation, MCP server status, and live SAP system connectivity. Reports findings with actionable fixes for each issue found.

<Response_Prefix> Every response triggered by this skill MUST begin with [Model: <main-model> · Dispatched: <sub-summary>] per ../../common/model-routing-rule.md § Response Prefix Convention. </Response_Prefix>

<Use_When>

• Something isn't working and you're not sure where the problem is
• After initial setup to confirm everything is configured correctly
• After a SAP system change (password reset, IP change, client migration)
• User says "doctor", "diagnose", "check connection", "something's broken", or "is it working"
• MCP tools return errors or are not visible in Claude Code </Use_When>

<Diagnostic_Checks> MANDATORY: Run the diagnostic checklist defined in diagnostic-checks.md. It covers Layer 1 (Plugin Health) → Layer 2 (MCP Server) → Layer 3 (SAP System Connection) → Layer 4 (Required ABAP Objects — gated on Layer 2 + 3 PASS) → Layer 5 (Configuration) → Layer 6 (RFC Backend, conditional) → Layer 7 (Cache Hygiene, always runs). Report PASS / FAIL / WARN / INFO per layer in the order defined there. </Diagnostic_Checks>

<Cache_Management> /sc4sap:sap-doctor accepts two cache flags passed via {{ARGUMENTS}}:

• --prune — after running the normal diagnostic report, invoke node "<plugin>/scripts/prune-cache.mjs" (dry-run) and render its output directly. No files are deleted.
• --prune --yes — after the normal report, invoke the script with --yes to actually delete every stale version directory.

Always show the ACTIVE version first and explicitly exclude it from deletion. If no stale versions exist, print "Cache is clean — nothing to prune." and return. If the script exits non-zero, surface the stderr verbatim in the doctor report and do NOT swallow the failure. </Cache_Management>

<Output_Format>

SC4SAP Doctor Report
====================
Plugin Health       [PASS]  v0.2.2 (cache matches marketplace)
MCP Server          [PASS]  plugin:sc4sap:sap responding, bridge preflight OK
SAP Connection      [PASS]  SID=S4H · Client=100 · User=PAEK
Required Objects    [WARN]  9a: 3/3 · 9b: 7/7 (ZCL_S4SAP_CM_ALV inactive)
Configuration       [PASS]
Cache Hygiene       [WARN]  3 stale versions consuming 1.7 GB (0.5.4, 0.6.0, 0.6.1)

Issues Found: 0 errors, 2 warnings

Fix: Activate ZCL_S4SAP_CM_ALV, or re-run /sc4sap:setup wizard step 9b
     /sc4sap:sap-doctor --prune --yes   (reclaim 1.7 GB)

Second example — connectivity failure gating Layer 4:

SC4SAP Doctor Report
====================
Plugin Health       [PASS]  v0.2.2
MCP Server          [FAIL]  plugin:sc4sap:sap not responding
SAP Connection      [SKIP]  Cannot test without MCP server
Required Objects    [SKIP]  SAP connection not ready
Configuration       [WARN]  No .sc4sap/config.json found
Cache Hygiene       [PASS]  0 stale versions

Issues Found: 1 error, 1 warning

Fix: Run /sc4sap:mcp-setup to install and register plugin:sc4sap:sap

</Output_Format>

<Remediation_Routing>

• Plugin version drift (cache ≠ marketplace) -> run /reload-plugins, then restart Claude Code
• MCP server missing -> /sc4sap:mcp-setup
• SAP connection error -> display connection troubleshooting from /sc4sap:mcp-setup Troubleshooting section
• Missing ZMCP_ADT_UTILS / ZMCP_ADT_DISPATCH / ZMCP_ADT_TEXTPOOL -> re-run /sc4sap:setup (wizard step 9a)
• Missing ZIF_S4SAP_CM / ZCX_S4SAP_EXCP / ZCL_S4SAP_CM_* -> re-run /sc4sap:setup (wizard step 9b)
• Any 9a/9b object inactive -> activate via ADT, or re-run the matching wizard step to reinstall source and activate
• Missing config -> run /sc4sap:setup wizard
• Authorization errors -> display required authorization objects (S_DEVELOP, S_TRANSPRT)
• Stale cache ≥ 500 MB -> /sc4sap:sap-doctor --prune to preview, then /sc4sap:sap-doctor --prune --yes to delete
• All pass -> "SC4SAP is healthy. System: {SID} Client: {client} User: {user}" </Remediation_Routing>