padosoft-knowhow-harvestlisted
Install: claude install-skill padosoft/skills
# Know-how harvest
Every repository you run accumulates hard-won rules — in a lessons file, a rules folder, decision records,
an internal skill. They stay there, benefiting one project. This is the loop that moves them into the shared
catalogue **without** re-reading everything every time, and without losing the record of what was decided.
**The rule: harvest the rule, leave the story, and write down what you decided — including the rejections,
because an undocumented rejection is re-mined every six months.**
---
## 0. Setup, once
```bash
cp .harvest-sources.example.json .harvest-sources.json # then fill in your paths
```
Both that file and `.harvest-ledger.json` are gitignored on purpose: **a list of the repositories you run is
metadata about your business**, and this repository is public. If more than one person harvests, keep them
in a private companion repository or a secret rather than committing them here.
## 1. The work order
```bash
make harvest # or: python3 scripts/harvest.py scan
python3 scripts/harvest.py scan --id monolith # one source
python3 scripts/harvest.py scan --json # to drive another tool
```
It prints four groups, and the current catalogue to match against:
| Group | What it means |
|---|---|
| **New** | never considered — the bulk on a first run, a trickle afterwards |
| **Changed** | reconsider; the work order says what you decided last time |
| **Gone** | the source file is gone — check whether a skill s