agentleFS
eugenezhangco-speceugenezhangco-spec/athena/.claude/skills/spending-summary/SKILL.md

spending-summary

Skill0 starsChanged 7 months agoMITView on GitHub

Query the user's Notion Expenses database and produce spending summaries — by month, category, merchant, tax-deductible status, or arbitrary date range. Use when the user says "how much did I spend", "summarize my spending", "budget for [period]", "tax-deductible total", "what did I spend on [category]", "monthly summary", "spending breakdown", or asks anything analytical about their expenses.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the keys aren't configured or the database has zero rows, say so and stop — don't fabricate numbers.

    line 27
  2. Very specific:

    Use mcp__notion__notion-fetch or the Notion data-source query tool to pull the matching rows.

    line 49
  3. Distinctive:

    If the dataset is small (under 200 rows), pull everything and aggregate locally for flexibility.

    line 59
  4. Very specific:

    If they asked for a monthly summary → offer to export to CSV for their accountant (mcp__filesystem__write_file, if filesystem MCP is …

    line 115
  5. Very specific:

    If they're over a mental budget → offer to flag the category in snapshot.md

    line 116
  6. Very specific:

    If they asked for tax-deductible → remind them to double-check with their accountant.

    line 117
  7. Distinctive:

    Never invent data. If the query returns zero rows or the database isn't configured, say so plainly.

    line 121
  8. Distinctive:

    Keep currencies separate in the output.

    line 122
  9. Distinctive:

    If the user asks about a category that doesn't exist in their data, say so and list the categories that do appear.

    line 124
  10. Distinctive:

    Never suggest tax strategies.

    line 125
  11. Very specific:

    After presenting numbers, Athena's coaching lens can surface:

    line 129
  12. Distinctive:

    Keep observations tight. One sentence.

    line 134
  13. Distinctive:

    Always lead with the numbers.

    line 138

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: Spending Summary

When to Use

  • User asks "how much did I spend [in some period / on some category]"
  • User asks for a monthly / weekly / quarterly / YTD summary
  • User asks "what's my tax-deductible total for [period]"
  • User asks "what's my biggest expense category this month"
  • User asks "how often do I spend at [merchant]"
  • User asks "am I over budget on [category]"
  • User asks for data to send to their accountant

Prerequisites

Reads from the user's Notion Expenses database. Same keys as log-expense:

expenses_database_id: <user's Notion database ID>
expenses_data_source: collection://<data source id>

If the keys aren't configured or the database has zero rows, say so and stop — don't fabricate numbers.

Expected Schema

Same as log-expense:

  • Name (title), Amount (number), Currency (select), Category (select), Merchant (text), Date (date), Payment Method (select), Tax Deductible (checkbox), Business (checkbox), Notes (text)

Flow

Step 1 — Parse the ask

Extract filters from the user's message:

  • Period — "this month", "last week", "Q1 2026", "2025", "March", "since March 1", "this year"
  • Category — any value from the Category select, or keyword matches ("food", "travel")
  • Merchant — specific name or pattern
  • Tax status — "tax-deductible", "business only", "personal"
  • Aggregation — total, average, breakdown by category, top merchants, frequency

Resolve relative dates to absolute dates in the user's timezone (from .claude/rules/operating-rules.md User Profile).

Step 2 — Query

Use mcp__notion__notion-fetch or the Notion data-source query tool to pull the matching rows. Typical filters:

User askNotion query
"How much this month"Date between first-of-month and today
"Food spending this year"Date ≥ Jan 1 2026, Category = "Food & Drink" OR "Groceries"
"Tax-deductible Q1"Date in Q1, Tax Deductible = true
"Biggest category this month"Date in month, group by Category, sum Amount
"How often at Starbucks"Merchant contains "Starbucks", count

If the dataset is small (under 200 rows), pull everything and aggregate locally for flexibility.

Step 3 — Aggregate

Compute the right shape for the question:

  • Total: sum of Amount, grouped by Currency (don't collapse different currencies).
  • Breakdown by category: sum + percent of total, sorted descending.
  • By merchant: top 5 merchants by spend + count.
  • Over time: bucket by week/month, show trend.
  • Tax-deductible: total of rows with Tax Deductible = true, with a note about consulting an accountant.

Step 4 — Report

Output format depends on the interface:

On Telegram — compressed, plain text, under 10 lines:

April 2026 spending:
Total: $1,847 USD
Food & Drink: $412 (22%)
Transport: $389 (21%)
Groceries: $285 (15%)
Shopping: $234 (13%)
Other: $527 (29%)

Tax-deductible: $178

In Claude Code — slightly richer, with a one-line observation at the bottom:

## April 2026 Spending

Total: $1,847 USD (12 days into the month)

| Category | Amount | % |
|---|---|---|
| Food & Drink | $412 | 22% |
| Transport | $389 | 21% |
| Groceries | $285 | 15% |
| Shopping | $234 | 13% |
| Software & SaaS | $187 | 10% |
| Business | $140 | 8% |
| Other | $200 | 11% |

Tax-deductible total: $178

Observation: Food & Drink is trending 18% higher than March. Mostly takeout.

Step 5 — Offer the next action

Based on what the user asked, suggest a follow-up:

  • If they asked for a monthly summary → offer to export to CSV for their accountant (mcp__filesystem__write_file, if filesystem MCP is enabled)
  • If they're over a mental budget → offer to flag the category in snapshot.md
  • If they asked for tax-deductible → remind them to double-check with their accountant; Athena is not a tax advisor

Rules

  • Never invent data. If the query returns zero rows or the database isn't configured, say so plainly.
  • Keep currencies separate in the output. Mixed-currency totals are usually wrong.
  • Round to the nearest unit in output; keep full precision in calculations.
  • If the user asks about a category that doesn't exist in their data, say so and list the categories that do appear.
  • Never suggest tax strategies. Surface the deductible total and stop.

Observations the mentor might add

After presenting numbers, Athena's coaching lens can surface:

  • Pattern detection: "You've spent $400 on food delivery this month. Last month was $180."
  • Underpricing pattern: if the user runs a business and expenses are low relative to stated revenue
  • Comfort-zone work: if a category keeps creeping up (e.g. retail therapy after stressful weeks documented in diary.md)

Keep observations tight. One sentence. Never moralize.

Output Format

Always lead with the numbers. Observations come last, one sentence, optional.

Related Skills

  • log-expense — capture a receipt/invoice into the Expenses DB
  • compile — synthesize spending history into a longer-term brief ("what are my financial patterns?")
  • weekly-review — budget audit fits naturally into the weekly review flow

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.