teamtinvio/jaz-ai / .agents/skills/jaz-kit
jaz-kit
VerifiedUse this skill when an accountant, bookkeeper, or owner is running real books in Jaz across one or more organizations from the terminal: setting up a workspace, switching between clients, running a month/quarter/year close that survives across sessions, or clearing a queue of drafts for review. Triggers on "open <org>", "close the books for June", "what needs my review", "what's due", "add another organization".
Jaz Kit
The operator layer for people who run books in Jaz. Where the other skills know how to do accounting, this one knows whose books these are, what was decided last time, and where the close left off.
The problem it solves
A close is not one conversation. Month-end runs eighteen steps over one to three days, and an accountant serving eight clients runs it eight times with eight different sets of bank accounts, materiality thresholds, and recurring accruals. Without somewhere to keep that, every session re-asks what it should already know, and every interruption loses the thread.
Jaz Kit gives each organization a folder: what the organization is, how it likes things done, what has been decided, and exactly where the current close stands.
The workspace
~/Documents/Jaz Kit/ override with JAZ_KIT_HOME
KIT.md who you are, index of organizations
_shared/policies/ skills/ firm-wide, applies to every organization
orgs/<slug>/
ORG.md the close config + session journal
.env optional API-key access only (gitignored)
policies/ rules/ skills/ how this organization works
scripts/ work/ automations · scratch (swept on exit, after asking)
closes/<period>/
CLOSE.md resumable state for this close
workpapers/ permanent audit file
ORG.md's organization_id identifies the company's ledger. Use OAuth by default: sign in with clio auth login, then pin every command with --org oauth:<organization_id>. CLI and local MCP share the local sign-in; tokens stay outside the kit. Hosted MCP uses its own OAuth connection and the same explicit organization ID. Verify that each connection returns that ID before work. Existing per-company .env API keys remain an optional route; see references/workspace.md.
Operations
Every flow is specified in references/flows.md; that file is the procedure, and Claude Code's /jk-* commands are thin pointers into it. In agents without slash commands, the trigger phrases below reach the same flows.
| Intent | Flow | Command |
|---|---|---|
| Set up the kit · add a company | setup | /jk-setup |
| Start work on one organization | open | /jk-open <org> |
| Run a period close | close | /jk-close <period> |
| Clear the draft review queue | review | /jk-review |
| See every organization at a glance | status | /jk-status |
| Connect, check, or rotate a key | keys | /jk-keys |
| Record how this organization works | policy | /jk-policy |
| Save a procedure for next time | teach | /jk-teach |
| Checkpoint mid-close | save | /jk-save |
| End the session cleanly | exit | /jk-exit |
| What can I do right now | help | /jk-help |
Rules that hold across every flow
Six ground rules govern every operation: one organization per session, draft first, verify before retrying a write, never print a key, record judgment not activity, and locking is the user's call. They are stated once, with their reasoning, at the top of references/flows.md. Read them before the first write of any session.
This is a harness, not a rulebook
Jaz Kit supplies structure: memory between sessions, sequencing, safety rails, and an audit trail. The accountant supplies the domain truth: their materiality, their filing dates, their accounting policies, their sign-off model, their procedures.
So: never assert a statutory deadline, a reporting standard, or a "best practice" from memory, and never offer a jurisdiction default for the user to correct. A confident wrong answer costs them more than an open question. When a fact is missing from ORG.md, ask for it and record it. When they tell you how they do something, write it to policies/ or rules/ and follow it thereafter.
The value is that they say it once.
What this skill does not do
It does not perform accounting. Close playbooks live in jaz-jobs, transaction patterns and calculators in jaz-recipes, payload shapes and error handling in jaz-api, migrations in jaz-conversion. When a flow reaches real work, it loads the relevant skill and follows it. If you find yourself about to write an accounting procedure into this skill, extend the right one instead and link to it.
Supporting files
- references/flows.md: the ground rules and the canonical procedure for every operation
- references/workspace.md: KIT.md, ORG.md, and
.envschemas; the auth model; sharing a kit across a team - references/close-state.md: CLOSE.md schema, the resume protocol, review-pack and close-summary formats
- references/org-interview.md: what to read automatically, and the four questions worth asking
- references/templates.md: copy-paste scaffolds for every file the kit creates
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |