feat: archive zoo_backup for home sync

This commit is contained in:
Patrick Plate
2026-06-24 19:27:05 +02:00
parent 02844e4c4a
commit 038e546963
133 changed files with 19953 additions and 0 deletions
@@ -0,0 +1,141 @@
---
name: cobol-lookup
description: Look up PAISY COBOL domain knowledge — PAI programs, DAI ISAM files, API interfaces, batch framework, ISAM tools (PDDI/Quidam), COKZ system, and COBOL architecture. Use when asked about a PAI program, DAI file, COBOL data flow, batch processing, or PAISY backend internals.
---
# Skill: cobol-lookup
Look up PAISY COBOL domain knowledge from the paisy-ai analysis repository.
## When to use
- User asks about a specific PAI program (e.g., "What does PAI022 do?")
- User asks about a DAI ISAM data file (e.g., "What fields are in DAI28?")
- User asks about a PAI API interface (e.g., "How does Java call PAI022?")
- User asks about COBOL architecture, COKZ system, batch framework, or ISAM tools
- User needs to understand data flow between COBOL programs
- User asks about wiki cross-check findings for specific Sachgebiete
## When NOT to use
- For Java module questions → use `domain-lookup` skill instead
- For Jira/Git operations → use `jira-lifecycle` or `create-worktree`
- For code changes → switch to Code mode
- For bulk indexing of all COBOL docs → use `cobol-index` skill instead
## Required Inputs
| Input | Source | Example |
|-------|--------|---------|
| `QUERY` | User question or topic | `"PAI022"`, `"DAI28 fields"`, `"batch framework"`, `"COKZ system"` |
| `CONTEXT` | Additional context (optional) | `"for ESIDEPAISY-12081"`, `"related to EAU"` |
## Output
- Structured answer with source file paths cited
- Key findings stored in BigMind (`category: "paisy-cobol"`)
## Steps
### 1. Parse query → determine lookup type
| Query pattern | Lookup type | Target path |
|--------------|-------------|-------------|
| `PAI<nnn>` or program name | PAI program | `analysis/PAI<nnn>.md` |
| `DAI<nn>` or ISAM file | DAI datamodel | `analysis/datamodel/DAI<nn>.md` |
| `api-PAI<nnn>` or "API for PAI..." | API interface | `analysis/api/api-PAI<nnn>.md` |
| Architecture, COKZ, structure | Architecture | `analysis/cobol-architecture.md`, `analysis/cokz-system.md` |
| Batch, payroll scripts, AWK | Batch framework | `analysis/batch-client/*.md` |
| PDDI, Quidam, ISAM tools | ISAM tools | `analysis/isam-tools/*.md` |
| Sachgebiet, wiki cross-check | Wiki cross-check | `analysis/wiki-crosscheck-*.md` |
| General overview | Overview | `CLAUDE.md` (repo root) |
### 2. Search BigMind first
```python
memory_search_facts("<program or topic name>")
memory_search_facts("<program or topic name> paisy-cobol")
```
If BigMind has a recent, complete answer → present it directly and skip to step 6.
### 3. Read the target doc from paisy-ai
```bash
cat /Users/pplate/git/paisy-ai/analysis/<target-file>.md
```
If the exact file doesn't exist, search for related content:
```bash
ls /Users/pplate/git/paisy-ai/analysis/ | grep -i "<keyword>"
grep -rl "<keyword>" /Users/pplate/git/paisy-ai/analysis/ --include="*.md" | head -10
```
### 4. Cross-reference related docs
After reading the primary doc:
- If it's a PAI program → check if an API doc exists: `analysis/api/api-PAI<nnn>.md`
- If it's a PAI program → check which DAI files it references (mentioned in the doc)
- If it's a DAI file → check which PAI programs use it (grep across program docs)
- If it's an API → read the corresponding PAI program doc for full context
### 5. Present findings structured
Format the response as:
```
**<Program/File Name>** — <one-line summary>
<Detailed explanation from the doc>
**Source:** `analysis/<path>.md`
**Related:** <list of related PAI/DAI/API docs if any>
**See also:** <ADP Wiki page if relevant>
```
### 6. Store key findings in BigMind
```python
memory_store_fact(
category="paisy-cobol",
fact="<program/file>: <concise summary of purpose, key fields, or behavior>"
)
```
Only store if the fact is new or significantly more detailed than what BigMind already has.
### 7. Fall back to ADP Docs Wiki
If paisy-ai docs don't cover the topic:
```python
# Check BigMind index for known Wiki pages
memory_search_facts("<topic> adpdocs")
# Search Wiki directly
set-wiki(uri="mcp://wikis/adpdocs.de.adp.com")
search-page(query="<COBOL topic keywords>")
```
Relevant Wiki pages for COBOL topics:
- Programmabläufe → Page ID 15201
- Einzelaufrufe BATCH → Page ID 15250
- Fehlermeldungen → Page ID 15196
- Umgebungsvariablen → Page ID 22767
## Error Handling
| Error | Resolution |
|-------|------------|
| Doc file not found | Search with `grep -rl` across analysis/; the program may have a variant name |
| BigMind returns nothing | Proceed to read from paisy-ai docs directly |
| paisy-ai repo not cloned | Run `git clone` or inform user the repo is needed at `/Users/pplate/git/paisy-ai` |
| Multiple matches | Present all matches, let user pick the most relevant |
| Doc is empty or stub | Note the gap, fall back to ADP Wiki, suggest running `cobol-index` to refresh |
## Language
- Answers: match the user's language (German or English)
- COBOL program names, field names, Sachgebiet codes: preserve as-is
- German payroll terms: use as-is (Lohnkonto, Fehlzeiten, Sachgebiet, etc.)