/sap-atc

name: sap-atc
description: |
  Runs the SAP ABAP Test Cockpit (ATC) end-to-end as a quality gate:
  builds an SCI Object Set scoped to the target object(s), creates an
  ATC Run Series bound to that set, polls the ATC Run Monitor until the
  run completes, then reads the Priority 1 / Priority 2 / Priority 3
  finding counts from the Manage Results screen and (best-effort)
  downloads the result text file. Applies the customer brief's
  MAX_PRIORITY threshold to decide pass / fail.
  Replaces the legacy SCI-results-tree implementation that lacked a
  Priority column. Per-stage VBS references are recorded against the
  S/4HANA 1909 ATC layout — re-record on first failure with
  /sap-gui-probe --record if your release uses different tree node IDs / grid
  column IDs.
  Prerequisites: Active SAP GUI session (use /sap-login first).
argument-hint: "<OBJECT_TYPE> <OBJECT_NAME> | --object-list=<file> [--variant=<NAME>] [--object-provider=<ID>] [--max-priority=<n>] [--object-set=<NAME>] [--run-series=<NAME>] [--poll-interval=<sec>] [--max-wait=<sec>] [--save-to=<PATH>] [--drill] [--no-drill]"

SAP ATC Quality Gate Skill

You drive the proper SAP ATC pipeline — Object Set → Run Series → Run Monitor → Manage Results — and apply a priority-based gate. The four stages each have their own VBS reference; SKILL.md orchestrates them with a polling loop in the middle.

Task: $ARGUMENTS

---

Stage map

Stage 1: SCI       /nSCI       create / refresh Object Set "<SET>"
   |
   v
Stage 2: ATC       /nATC       tree node 12 → CREATE_SERIE bound to <SET>
   |                            → EXECUTE_SERIE → run is async
   v
Stage 3: ATC       /nATC       tree node 13 → poll Run Monitor for <RUN>
   |                            until STATE=COMPLETED (or FAILED)
   v
Stage 4: ATC       /nATC       tree node 14 → Manage Results: read
                                P1/P2/P3 counts; best-effort download
                                to <OUTPUT_PATH>
   |
   v
Gate logic in SKILL.md compares the counts to MAX_PRIORITY.
   |
   v
Stage 4b: ATC      /nATC       (optional) tree node 14 → drill into the
                                run-series row → export per-finding ALV
                                as TSV. Auto-triggered when gate FAILS;
                                always triggered by --drill. Outputs
                                <OUTPUT_PATH>.findings.tsv with one row
                                per finding: PRIO | CHECK_ID | CHECK_TITLE
                                | OBJ_NAME | OBJ_TYPE | MSG_TEXT | LINE.
   |
   v
Final PASS / FAIL emitted with the findings.tsv path when available.

---

Shared Resources

| File | Stage | Purpose | |---|---|---| | `<SAP_DEV_CORE_SHARED_DIR>/rules/skill_operating_rules.md` | — | Mandatory operating rules | | `<SAP_DEV_CORE_SHARED_DIR>/rules/language_independence_rules.md` | — | GUI-scripting language independence — identify by component ID + DDIC field name, status-bar checks via `MessageType` codes (S/W/E/I/A), VKey instead of menu-text, no branching on `.Text`/`.Tooltip`/window titles | | `<SAP_DEV_CORE_SHARED_DIR>/rules/sap_gui_security_handling.md` | 4 / 4b | SAP GUI Security dialog handling — the Stage-4 result **download** and Stage-4b findings-TSV **export** are SAP-GUI-side file IO, so they can raise the modal "SAP GUI Security" dialog (which suspends the Scripting API and hangs cscript). Pre-check + OS-level watcher wrap both downloads. | | `<SAP_DEV_CORE_SHARED_DIR>/scripts/sap_gui_security_precheck.ps1` | 4 / 4b | Read-only allow-list pre-check (`saprules.xml`) — `ALLOWED` (exit 0) / `NOT_COVERED` (exit 1). | | `<SAP_DEV_CORE_SHARED_DIR>/scripts/sap_gui_security_sidecar.ps1` | 4 / 4b | OS-level (Win32) watcher that auto-dismisses the SAP GUI Security dialog (ticks Remember + clicks Allow). Launched as a background process before each download. | | `<SAP_DEV_CORE_SHARED_DIR>/scripts/sap_object_resolver.ps1` | 2b | Pre-flight existence check of the target (CLI mode, 32-bit PowerShell; emits `STATUS: RESOLVED\|NOT_FOUND\|AMBIGUOUS\|UNKNOWN_TYPE\|RFC_ERROR`). `NOT_FOUND` aborts the skill with `ATC_EMPTY_SCOPE` before any scope is built. | | `<SAP_DEV_CORE_SHARED_DIR>/scripts/sap_artifact_lib.ps1` | 6d | `New-SapScopeKey` / `Register-SapArtifact` — register the result TXT + findings TSV so `/sap-evidence-pack` collects them under the object's scope (Kind `atc_findings`). Best-effort; never changes the gate verdict. | | `<SKILL_DIR>/references/sap_sci_create_object_set.vbs` | 1 | Define the scope — programs / classes / FMs / interfaces / FUGRs | | `<SKILL_DIR>/references/sap_atc_create_run_series.vbs` | 2 | Schedule + trigger the ATC run | | `<SKILL_DIR>/references/sap_atc_check_run_status.vbs` | 3 | Read run state from the Monitor (read-only; safe to call in a poll loop) | | `<SKILL_DIR>/references/sap_atc_get_results.vbs` | 4 | Pull P1/P2/P3 counts + try to save the result TXT (outer grid only — summary level) | | `<SKILL_DIR>/references/sap_atc_drill_findings.vbs` | 4b | (Optional) Drill into the run-series row → export per-finding ALV as TSV. Run when gate FAILS or `--drill` is passed. | | `<SAP_DEV_CORE_SHARED_DIR>/scripts/sap_error_hints.ps1` | 6c | frequently_errors recorder. `-Action record -Source ATC -FindingsFile <...>.findings.tsv` captures FM/METHOD-attributable findings to the team store as CANDIDATEs. Best-effort; never changes the gate verdict. |

---

Step 0 — Resolve Work Directory

**Resolve `work_dir` via the env-aware helper** — do NOT take `work_dir` from a direct `settings.json` read (that ignores the `SAPDEV_AI_WORK_DIR` env var and `userconfig.json`). Use the `WORK_DIR=` value printed by:

powershell -NoProfile -ExecutionPolicy Bypass -Command ". '<SAP_DEV_CORE_SHARED_DIR>\scripts\sap_settings_lib.ps1'; . '<SAP_DEV_CORE_SHARED_DIR>\scripts\sap_connection_lib.ps1'; Write-Output ('WORK_DIR=' + (Get-SapWorkDir)); Write-Output ('RUN_TEMP=' + (Get-SapRunTemp))"

The settings note below still applies