ForemanStack cards
It knows your stack, and costs no context until it matters
Each card gives the commands that verify the work, a failing-test recipe, the mistakes models actually make and a review checklist. Cards cost nothing until a matching file is edited.
Show the numbers
| value | |
|---|---|
| python | 62 |
| data | 60 |
| infra | 60 |
| tools | 59 |
| security | 59 |
| design | 58 |
| web | 57 |
| node | 56 |
| cloud | 55 |
| ml | 54 |
| lang | 50 |
| testing | 50 |
| jvm | 50 |
| domains | 45 |
| native | 41 |
| frontend | 37 |
| mobile | 28 |
| claude-code | 27 |
| rust | 26 |
| practice | 25 |
| ops | 24 |
| go | 24 |
| dotnet | 22 |
| games | 21 |
| idioms | 16 |
| docs | 10 |
| hardware | 8 |
Picked by what you touch
fm lib picks the cards that match your repo's manifests and files. The first edit of a file a card covers brings in that card's pitfalls, once per session; a file that imports a framework gets that framework's card too. fm lib source add indexes another skills repo read-only.