finance-setup
Skill2 starsChanged 51 days agoMITView on GitHub
Set up a private personal finance workspace, shared calculation runtime, and gradual account and household intake. Use for initial setup or missing prerequisites, not for routine investment research.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- finance-dashboardSkill
- finance-gatherSkill
- finance-memorySkill
- finance-planSkill
- investment-researchSkill
- portfolio-reviewSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 9 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈9
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Create a usable private workspace and explain what is known, missing, and connected.
- Distinctive: line 9
Start with the user's immediate question.
- Distinctive: line 10
Run installation and calculation commands yourself with the host's tools.
- Distinctive: line 12
Ask for help only when a required capability, permission, or login is unavailable.
- Distinctive: line 14
Before collecting personal financial records, recommend turning off optional sharing for model training or product improvement in the …
- Distinctive: line 16
Ask the user to check their agent and every model provider it uses, including optional research tools.
- Very specific: line 25
Read the active project's applicable AGENTS.md and the selected workspace's setup.md, when present, for integration settings.
- Very specific: line 38
Use --demo only when explicitly requesting synthetic demonstration data in a separate workspace.
- Very specific: line 40
Read the initialized files and run finance-core validate --workspace /private/path.
- Distinctive: line 45
Read the intake guide when collecting profile or account facts.
- Distinctive: line 49
Start with account coverage, the planning question, and desired reporting currency.
- Distinctive: line 64
Prefer read APIs, then downloaded exports, then user-authorized computer use.
- Very specific: line 68
Keep limits in rules.json unset unless the user supplies them.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Finance setup
Create a usable private workspace and explain what is known, missing, and connected. Start with the user's immediate question; do not require a complete financial biography. Run installation and calculation commands yourself with the host's tools. Keep the conversation natural; do not hand the user a Python setup checklist. Ask for help only when a required capability, permission, or login is unavailable.
Before collecting personal financial records, recommend turning off optional sharing for model training or product improvement in the provider's privacy or data controls. Ask the user to check their agent and every model provider it uses, including optional research tools. Use current official guidance for the exact settings. Explain that the provider still processes requests and may retain data even when training is disabled.
Establish the runtime and workspace
- Reuse the user's existing workspace and configured environment when known.
Otherwise select a private directory outside the public source checkout and confirm
the base currency. A new currency choice does not imply residence or nationality.
Read the active project's applicable
AGENTS.mdand the selected workspace'ssetup.md, when present, for integration settings. If the project supplies account, research, or dashboard adapters, follow project integration. - Reuse a working
finance-corecommand when available. Otherwise follow runtime setup to obtain the official source and install its helpers in an isolated environment. A global skills install may contain only skill folders, so a missing source checkout is expected. You can fetch it when needed. Record the working runtime path for later skills; do not modify global PATH or unrelated environments merely to make a bare command available. That reference also covers diagnosing an existing setup and updating skills, the runtime, or optional engines without replacing private records. - Initialize with
finance-core init --workspace /private/path --base-currency USD, replacing the illustrative path and currency with the user's choices. Use--demoonly when explicitly requesting synthetic demonstration data in a separate workspace. Demonstration records are never real-account fallbacks. - Read the initialized files and run
finance-core validate --workspace /private/path. Record no passwords, recovery codes, authentication cookies, or API secrets in them.
Interview gradually
Read the intake guide when collecting profile or account facts. Reuse established answers and ask only questions that affect the current task. Let the user skip sensitive or unavailable information; skipped fields remain unknown.
Start with account coverage, the planning question, and desired reporting currency. Then collect relevant income, expense, household, goals, and tax-jurisdiction facts. Distinguish what the user said from an inference and from a researched legal conclusion. Neither location nor payroll alone establishes citizenship or tax residence.
facts.json and decisions.json use append-only records with user confirmation and
sources. To save intake statements, use the envelope in the intake guide and
finance-core memory --workspace /private/path --input /private/path/intake.json.
Save explicit answers within the authorized intake without asking for the same
confirmation again. Leave unanswered questions in a private intake note.
Configure only what is needed
Inventory institutions using stable aliases such as main-brokerage; real account
numbers belong only in private connector settings if required by a provider.
Prefer read APIs, then downloaded exports, then user-authorized computer use.
Initial authorization may require the user's own login and multifactor authentication.
Store secrets in an available keychain or secret manager, or environment variables.
Keep limits in rules.json unset unless the user supplies them. A null limit means
report only. Keep both model-sharing switches false until the user authorizes that
specific export. Export permission does not authorize arbitrary model transmission.
Explain that a hosted agent may send files it reads and pages it sees to its model
provider. Local storage and disabled context exports do not make that inference local.
Optional research engines and connectors are unnecessary for an initial ledger.
Install them only when the requested workflow needs them. Before installing, check
current official documentation for supported authentication and data coverage.
Handoff
Save the runtime command and any selected project adapters in the private workspace's
setup.md for later agent sessions. Tell the user where their records live, the base
currency, and which accounts are covered.
Describe accounts as connected, awaiting export, or unknown; an empty ledger is not
a zero net worth. Identify the next useful missing input rather than a long questionnaire.
The workspace should contain ledger.json, transactions.json, lots.json,
facts.json, rules.json, and decisions.json; reports and evidence stay private too.
State whether validation passed, and never claim setup implies any account was synced.
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.

