name: personal-finance-skill description: > Personal finance management skill providing 75 tools across 7 extensions for bank account aggregation (Plaid), brokerage trading (Alpaca), portfolio monitoring (IBKR), tax optimization, market intelligence (Finnhub, SEC EDGAR, FRED, BLS, Alpha Vantage), social sentiment (StockTwits, X/Twitter, Quiver), and financial analysis. Supports scheduled scans, anomaly detection, net worth tracking, tax-loss harvesting, and approval-gated execution of financial actions. license: MIT metadata: author: contributors version: "2.0.0" tools: "75" extensions: "finance-core, plaid-connect, alpaca-trading, ibkr-portfolio, tax-engine, market-intel, social-sentiment"

Personal Finance Skill

A comprehensive personal finance management skill with 75 tools across 7 extensions for banking, investing, tax, market intelligence, social sentiment, and financial analysis workflows.

When to Use

Activate this skill when a user asks for:

• Account aggregation — connecting bank accounts, viewing balances, syncing transactions
• Net worth / cash flow — computing totals, tracking spending, savings rate analysis
• Portfolio monitoring — positions, allocation, performance, drift detection
• Trading — placing/canceling orders, market data, asset lookup (Alpaca)
• Tax optimization — estimated liability, TLH candidates, wash sale checks, quarterly payments
• Tax document processing — parsing W-2, 1099-B/DIV/INT, K-1, Form 1040, Schedules A-E/SE, Form 8949, Form 6251 (AMT), state returns
• Market intelligence — company news, SEC filings, analyst recommendations, economic data (FRED, BLS), news sentiment
• Social sentiment — StockTwits sentiment, X/Twitter cashtag analysis, trending symbols, congressional trading
• Recurring expense tracking — subscriptions, bills, income streams
• Anomaly detection — unusual transactions, balance drops, duplicate charges
• Financial briefings — weekly/monthly summaries with action items
• Scheduled finance workflows — cron-based scans, alerts, reports

Architecture Overview

Seven extensions organized in three layers:

Intelligence Layer
  tax-engine (23 tools) — parsing (15), liability, TLH, wash sales, lots,
    Schedule D computation, state tax, AMT
  market-intel (10 tools) — news, fundamentals, SEC filings, economic data
  social-sentiment (6 tools) — StockTwits, X/Twitter, congressional trades

Data Source Adapters
  plaid-connect (8)   alpaca-trading (10)   ibkr-portfolio (9)

Foundation Layer
  finance-core (9 tools) — canonical models, storage, normalization,
    policy checks, anomaly detection, briefs

Data flow: Adapters fetch provider data → finance-core normalizes and stores → intelligence layer analyzes → policy engine gates actions.

Tool Catalog

finance-core — 9 tools

Full schemas: references/ext-finance-core.md

plaid-connect — 8 tools

Full schemas: references/ext-plaid-connect.md

alpaca-trading — 10 tools

Full schemas: references/ext-alpaca-trading.md

ibkr-portfolio — 9 tools

Full schemas: references/ext-ibkr-portfolio.md

tax-engine — 23 tools

Full schemas: references/ext-tax-engine.md

market-intel — 10 tools

Full schemas: references/ext-market-intel.md

social-sentiment — 6 tools

Full schemas: references/ext-social-sentiment.md

Key Workflows

1. Onboarding — Connect Accounts

plaid_create_link_token(products: ["transactions", "investments", "liabilities"])
  → User completes Plaid Link
  → plaid_exchange_token(publicToken)
  → plaid_get_accounts → finance_upsert_snapshot(source: "plaid")
  → plaid_get_transactions → finance_upsert_snapshot
  → plaid_get_investments → finance_upsert_snapshot
  → finance_get_net_worth → present baseline to user

2. Daily Scan — Anomaly Detection

plaid_get_transactions(cursor) → finance_upsert_snapshot
alpaca_list_positions → finance_upsert_snapshot(source: "alpaca")
ibkr_auth_status → ibkr_get_positions → finance_upsert_snapshot(source: "ibkr")
  → finance_detect_anomalies(lookbackDays: 7)
  → Alert on medium/high severity findings

3. Tax-Loss Harvesting

finance_get_state(include: ["positions"])
  → tax_find_tlh_candidates(positions, marginalRate)
  → tax_check_wash_sales(proposedSales, recentPurchases)
  → tax_lot_selection(symbol, qty, lots)
  → finance_policy_check(actionType: "tax_move")
  → [If approved] alpaca_create_order(side: "sell", ...)

4. Quarterly Tax Review

tax_parse_w2 + tax_parse_1099b + tax_parse_1099div + tax_parse_1099int
  → tax_estimate_liability(filingStatus, income)
  → tax_quarterly_estimate(projectedIncome, priorYearTax, paymentsMade)
  → finance_generate_brief(period: "quarterly")

5. Portfolio Monitoring

alpaca_list_positions + ibkr_get_positions
  → finance_upsert_snapshot (both sources)
  → ibkr_portfolio_allocation (check drift)
  → alpaca_portfolio_history (performance trend)
  → finance_detect_anomalies
  → finance_generate_brief(period: "weekly")

6. Company Research — Market Intelligence

intel_company_news(symbol: "AAPL", limit: 10)
  → intel_analyst_recommendations(symbol: "AAPL")
  → intel_stock_fundamentals(symbol: "AAPL", freq: "quarterly")
  → intel_sec_filings(symbol: "AAPL", formType: "10-K")
  → intel_news_sentiment(tickers: "AAPL")
  → social_stocktwits_sentiment(symbol: "AAPL")
  → social_x_cashtag(symbol: "AAPL")

7. Economic Overview

intel_fred_series(seriesId: "GDP")
  → intel_fred_series(seriesId: "CPIAUCSL")
  → intel_fred_series(seriesId: "UNRATE")
  → intel_bls_data(seriesIds: ["CES0000000001"])
  → intel_fred_series(seriesId: "DFF")

8. Full Tax Return Processing

tax_parse_1040(rawData) → tax_parse_schedule_a(rawData)
  → tax_parse_schedule_b(rawData) → tax_parse_schedule_c(rawData)
  → tax_parse_schedule_d(rawData) → tax_parse_schedule_e(rawData)
  → tax_parse_form_8949(rawData) → tax_parse_schedule_se(rawData)
  → tax_compute_schedule_d(gains, losses, carryovers)
  → tax_compute_state_tax(stateCode, taxableIncome, filingStatus)
  → tax_compute_amt(taxableIncome, adjustments, regularTax)
  → tax_parse_state_return(rawData)

9. Congressional Trading Signals

social_quiver_congress(daysBack: 30)
  → Filter for large purchases
  → intel_company_news(symbol: <top_ticker>)
  → alpaca_market_data(symbols: <top_ticker>)
  → social_stocktwits_sentiment(symbol: <top_ticker>)

Configuration

Environment Variables

Extension Config

Each extension has an openclaw.plugin.json with a configSchema. Key settings:

• finance-core: storageDir, anomalyThresholds, policyRulesPath
• plaid-connect: plaidEnv, webhookUrl, clientName, countryCodes
• alpaca-trading: env (paper/live), maxOrderQty, maxOrderNotional
• ibkr-portfolio: baseUrl, defaultAccountId
• tax-engine: defaultFilingStatus, defaultState, defaultTaxYear
• market-intel: finnhubApiKeyEnv, fredApiKeyEnv, blsApiKeyEnv, alphaVantageApiKeyEnv, secEdgarUserAgent
• social-sentiment: xApiBearerTokenEnv, quiverApiKeyEnv

Cron Examples

Weekly Financial Brief

openclaw cron add \
  --name "Finance Weekly Brief" \
  --cron "0 8 * * 1" \
  --tz "America/Los_Angeles" \
  --session isolated \
  --message "Run personal-finance-skill weekly workflow: sync all providers, compute net worth delta, top spend changes, upcoming bills, tax posture, and portfolio drift. Send concise brief with action queue."

Daily Anomaly Scan

openclaw cron add \
  --name "Finance Daily Anomaly" \
  --cron "15 7 * * *" \
  --tz "America/Los_Angeles" \
  --session isolated \
  --message "Sync latest transactions, run finance_detect_anomalies. Alert on medium/high/critical findings only."

Quarterly Tax Check

openclaw cron add \
  --name "Quarterly Tax Review" \
  --cron "0 9 1 1,4,6,9 *" \
  --tz "America/Los_Angeles" \
  --session isolated \
  --message "Run quarterly tax review: estimate liability, check withholding gap, find TLH opportunities, assess quarterly payment risk."

Portfolio Drift Monitor

openclaw cron add \
  --name "Portfolio Drift Monitor" \
  --cron "*/30 13-21 * * 1-5" \
  --tz "America/New_York" \
  --session isolated \
  --message "Check portfolio allocation vs target bands. Alert if drift exceeds threshold for 2 consecutive scans."

Non-Negotiable Guardrails

These rules apply to all AI agents using this skill:

1. Always run finance_policy_check before any side-effecting action (trades, transfers, tax moves).
2. Never bypass approval requirements. If policy requires user or advisor approval, halt and request it.
3. Numeric outputs must come from deterministic calculators. Never use LLM arithmetic for tax amounts, P/L, or net worth — always use the tool.
4. Recommendations must include assumptions and data freshness. Every financial recommendation states what data it used and when that data was last updated.
5. Never expose raw access tokens or API keys in tool outputs or conversation.
6. Never auto-execute in live trading without explicit user confirmation, even if policy rules allow it.
7. All investment-related outputs must include disclaimer: "This is informational only, not financial advice. Consult a qualified advisor before making financial decisions."
8. If data is stale, say so. Report data freshness before advising.

Reference Index

Detailed documentation is available in the references/ directory: