folio-bokforing
Skill0 starsChanged 27 days agoApache-2.0View on GitHub
Complete (bokføre) Folio bank events after the user explicitly says yes — mark events complete or undo it, set categories and receipts without the folio-kladd guards, plus a write escape hatch. Opt-in; install only if you want the agent to book in Folio. Never moves money.
In this repository6 other files it gives its agents.
- fiken-bokforingSkill
- fiken-kladdSkill
- fikenSkill
- folio-kladdSkill
- folioSkill
- reconcileSkill
The file
Files
Folio bokføring (opt-in)
Completing an event marks it as booked. With Folio's Fiken integration on, it is also booked in Fiken, where booked entries can't be deleted.
The rule: run a command from this skill only after the user has said yes to that exact event in the same conversation. Show them the event first (date, amount, merchant, category, receipt) using the folio skill. Never complete events as part of a chain of other steps.
What locks this, honestly: nothing deterministic. Every command requires --confirm, but you can type that flag yourself. The real check is the user's agent tool asking before it runs a command containing --confirm. If the user hasn't set that up, recommend completing events in Folio instead.
Money stays out of reach regardless: raw refuses writes to /payments, /salary and /integration, and a payment created through Folio's API can't be sent without a person signing it in Folio.
F=~/.claude/skills/folio-bokforing/scripts/folio-bokforing.sh
bash "$F" complete <eventId> --confirm
bash "$F" uncomplete <eventId> --confirm
bash "$F" patch-event <eventId> --category <id> --confirm # no kladd guards
bash "$F" attach <eventId> <file> --confirm # no kladd guards
bash "$F" raw <METHOD> <path> [json] --confirmSetup: same FOLIO_WRITE_KEY as folio-kladd. Every call is logged to ~/.local/state/fiken-folio-skills/audit.jsonl.
Part of fiken-folio-skills by Fredrik Skorge Salberg (Pifre AS). Not affiliated with Folio.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.