---

name: dump-triage description: Triage ABAP short dumps (ST22) — list recent dumps, group by root cause, read the dump chapters, pull the source at the termination point, and produce a prioritized report with a concrete fix suggestion per dump family. Use whenever the user mentions dumps, ST22, runtime errors, CX_ exceptions, "production crashed", TYPELOAD errors, "why did this ABAP terminate", or asks for a morning system-health check. Works fully in read-only mode — safe to run against production.

Dump Triage

Turn a pile of ST22 short dumps into a short, prioritized action list. The workflow is read-only end to end, so it is safe to point at production systems.

Prerequisites

• sap-adt-mcp ≥ 0.8.1 connected to the target system (adt_ping to verify).
• NetWeaver 7.50+ or S/4HANA recommended — the ADT runtime-dumps feed (/sap/bc/adt/runtime/dumps) is not exposed on older releases. If adt_list_dumps 404s, the system doesn't serve the feed; tell the user to use ST22 directly and stop.
• SAP user authorizations: runtime-dump display (typically S_ADT / basis display roles). No write authorizations needed.
• Read-only mode: fully supported. This skill never writes. If a fix is agreed and the user wants it applied, that's a separate (write-mode) edit via adt_set_source — confirm explicitly before leaving read-only territory.
• Date filters use YYYYMMDD on on-prem systems (e.g. 20260611).

Workflow

1. Collect

adt_list_dumps with a sensible window — default to the last 24h unless the user says otherwise (from: <yesterday YYYYMMDD>). Filter by user only when asked; note the user filter is enforced client-side. Respect maxResults (default 20; raise for a weekly review).

2. Group before reading

Don't fetch every dump's detail. Group the list by (runtimeError, program) — that pair defines a dump family. 200 TYPELOAD_NEW_VERSION entries from one job are one problem, not 200. Rank families by:

1. frequency (count in window),
2. blast radius (PRD > QAS > DEV; batch users hitting it every minute),
3. severity of the error class (data-loss-ish errors like CONVT_NO_NUMBER, ITAB_DUPLICATE_KEY in update tasks rank above one-off TIME_OUTs).

3. Deep-read the top families

For the top 3–5 families: adt_get_dump with the family's most recent dump id. The default chapter set (shortText, whatHappened, errorAnalysis, howToCorrect, whereTerminated, sourceCodeExtract) is usually enough; request more chapters only when the analysis is inconclusive.

From whereTerminated, extract program/include + line. Then pull real context: adt_get_source for that include around the failing line (use firstLine/lastLine — ±40 lines is plenty; never fetch a 20k-line program whole). The dump's own sourceCodeExtract is narrow; the live source shows the surrounding logic, and whether the code has changed since the dump (adt_list_versions / adt_compare_versions when staleness is suspected).

4. Diagnose per family

State a root-cause hypothesis grounded in the chapters + source, not generic advice. Typical patterns worth checking explicitly:

• TYPELOAD_NEW_VERSION / LOAD_PROGRAM_*: something was transported or activated while jobs were running — correlate with adt_list_transports (recently released) rather than blaming the code.
• CX_SY_OPEN_SQL_DB / SQL errors: read the failing statement; if data dependent, a targeted adt_read_table SELECT can confirm the offending rows (read-only, requires NW 7.55+).
• CONVT_, COMPUTE_: almost always unvalidated input — find the field's origin with the source context.
• Untyped CX_ROOT in custom code: check whether the raise site is in a Z object (adt_where_used helps trace callers).

5. Report

One block per family, ranked:

DUMP TRIAGE — DEV, last 24h (37 dumps → 4 families)

#1  TYPELOAD_NEW_VERSION — 29× — program SAPLZSD_PRICING — users: BATCH
    Cause: TR E4DK900812 released 02:14, jobs ZPRICE_RECALC were mid-run.
    Fix: none needed in code; reschedule the job after imports, or use
    server-group restart. Prevention: import window outside job schedule.

#2  CONVT_NO_NUMBER — 5× — ZRPT_MARGIN line 214
    Cause: CONV #( lv_input ) on a field fed from file upload; row 3 of the
    file contains '12,5' (comma decimal).
    Fix: validate/replace decimal separator before CONV — patch sketch below.
    …

Close with: which families are ignorable (one-offs, already-fixed code, test-system noise) — saying what NOT to chase is half the value of triage. Offer follow-ups: "want me to draft the fix for #2?" (write mode) or "schedule this as a daily morning check".