菜单
Use when checking SAP system health, investigating reported errors, or performing a periodic operational health check. Provides a standardized workflow using ListDumps/GetDump to detect ABAP short dumps and route new findings into /triage for investigation.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/5throck/ai-workspace-standards --skill dump-monitoring命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Use when reviewing ABAP source for Clean ABAP conformance before object activation, transport release, or after importing foreign code. Covers the three ABAP-specific dimensions that general-purpose review skills do not - naming conventions, pretty-printer formatting, and anti-pattern compliance - against docs/clean-abap-checklist.md. Trigger on "review the ABAP code", "Clean ABAP check", "naming convention review", "pretty printer check", "anti-pattern scan".
Use when working on SAP ABAP development tasks — provides specialized workflows for BAPI exploration, transport management, unit testing, performance analysis, impact architecture analysis, and documentation audits. Trigger on any SAP/ABAP coding, debugging, or system analysis task.
Manual Post-Write QA chain for Claude Code Desktop App (hooks don't fire)
正在显示 SKILL.md
| name | Dump Monitoring |
| description | Use when checking SAP system health, investigating reported errors, or performing a periodic operational health check. Provides a standardized workflow using ListDumps/GetDump to detect ABAP short dumps and route new findings into /triage for investigation. |
| version | 1.0.0 |
| last_reviewed | 2026-08-15T00:00:00.000Z |
| status | active |
| scope | co-abap |
| owner | devops-admin |
| prerequisites | vsp MCP server |
| metadata | {"type":"core","triggers":["dump-monitor","ListDumps","GetDump","short dump","system health"]} |
Closes an observability gap: without this skill, ABAP short dumps are only found reactively
when a user reports an error. Owned by DevOps/Admin (agents/devops-admin.md); any agent
may run it as part of a health check.
1. List recent dumps
ListDumps(since=<last-check-timestamp or 24h ago>)
2. For each new dump not already triaged (cross-reference against
scratch/qa-reports/ or memory/ logs from prior checks):
GetDump(dump_id) → capture: program, exception, call stack, timestamp, user/client
3. Classify
- Runtime error in a Z*/custom object → candidate for /triage with classification "Debug"
- Runtime error in an SAP standard object → note only, do not modify standard code
- Repeated dump (same program/exception, multiple occurrences) → escalate priority
4. Route
- New, actionable dump → run `/triage "Investigate dump: <program> — <exception>"`
- Already-tracked dump → skip (avoid duplicate task files)
5. Record
Append a Dump Monitoring Report (below) to the current session's memory/YYYY-MM-DD.md
### Dump Monitoring — <date/time of check>
**Window checked**: <since timestamp> → now
**Dumps found**: <N> (<M> new, <N-M> previously triaged)
| # | Program | Exception | Timestamp | Occurrences | Action |
|---|---------|-----------|-----------|--------------|--------|
| 1 | ZCL_EXAMPLE | UNCAUGHT_EXCEPTION | 2026-07-10 14:02 | 3 | Routed to /triage |
**Verdict**: <Clean / N actionable () >
This harness has no built-in cron. To run this check periodically, wire an external scheduler (OS cron / Windows Task Scheduler / a CI scheduled workflow) to invoke the AI tool with a fixed prompt such as "Run the Dump Monitoring workflow from skills/dump-monitor/SKILL.md" — this keeps the mechanism portable across Claude Code, Gemini CLI, and Antigravity rather than tying it to one platform's session-level scheduler.
This skill provides a standardized workflow for detecting and triaging ABAP short dumps in an SAP system. Without it, dumps are only found reactively when users report errors. The workflow uses ListDumps and GetDump to proactively identify runtime errors and route actionable findings into the triage process for investigation and resolution.
ListDumps with a time window (default: last 24 hours or last check timestamp).GetDump to capture program, exception, call stack, timestamp, and user/client./triage for investigation; already-tracked dumps are skipped to avoid duplicates.memory/YYYY-MM-DD.md.A structured Dump Monitoring Report appended to the session memory log, containing: