agentleFS
alebgl77alebgl77/claude-inc/skills/reconciliation/SKILL.md

reconciliation

Skill21 starsChanged 4 months agoMITView on GitHub

Matches two datasets — bank vs ledger, subledger vs GL — via exact matching, fuzzy matching (amount tolerance, date window, reference similarity), and split detection; buckets exceptions (timing, missing, duplicate, amount mismatch), ages unmatched items, proposes adjusting entries, and proves the bridge between the two balances. Use when the user says 'reconcile the bank statement', 'these two reports don't agree', or 'the AR subledger is off from the GL'.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Propose adjusting entries: only for genuine book errors (unbooked bank fee, duplicate posting) — full balanced journals in the …

    line 26

“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “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

Reconciliation — Reconciler

"Reconcile the books."

When to use

  • Bank vs ledger — "reconcile June's bank statement against GL cash".
  • Subledger vs GL — "the AR subledger says 412k, the GL says 405k — why?".
  • Processor vs books — "match Stripe payouts to booked revenue".
  • Intercompany — "our payable to the sub doesn't equal their receivable from us".
  • Any two lists that should agree and don't.

Workflow

  1. Load both sides with code (pasted data or CSV/XLSX — never eyeball long columns). Standardize columns (date, amount, reference, description), normalize signs and currency, drop exact re-export duplicates. Record each side's row count and total — these anchor the bridge.
  2. Exact pass. Match on amount + date (+ reference when present); remove matched pairs from the pool.
  3. Fuzzy pass, rules tightening in order: amount within tolerance (default ±0.01 — widen only with user sign-off), date within window (default ±3 business days), reference/description similarity. Log every fuzzy match with the rule that made it.
  4. One-to-many pass. Detect splits and batches (one bank deposit = several ledger receipts) via combination sums inside the date window.
  5. Bucket every leftover: timing (legit, will clear — deposits in transit, outstanding checks), missing (no counterpart expected — unbooked fees, unrecorded receipts), duplicate (same item twice on one side), amount mismatch (paired but different — show the delta).
  6. Age the unmatched by days outstanding: 0–7, 8–30, 31–60, 60+. Anything over 30 days gets a priority flag.
  7. Propose adjusting entries only for genuine book errors (unbooked bank fee, duplicate posting) — full balanced journals in the journal-entry skill's format. Timing items never get forced adjustments.
  8. Prove the bridge: side A balance ± each exception bucket = side B balance, computed in code, residual shown. Target 0.00; anything else is stated, never hidden.

Output format

# Reconciliation — <A> vs <B> — <period>

## Bridge
<A> balance                     <amount>
  + timing: <description>       <amount>
  − missing on <side>: <desc>   <amount>
  ± mismatches (net)            <amount>
= <B> balance                   <amount>
Residual: 0.00 ✓  (or: <amount> — UNRESOLVED, see exceptions)

## Match summary
Exact: <n> | Fuzzy: <n> (rules logged) | One-to-many: <n> | Unmatched: <n>

## Exceptions
| # | Side | Date | Amount | Ref | Bucket | Age (days) | Action |
|---|------|------|-------:|-----|--------|-----------:|--------|
| 1 | Bank | <date> | <amt> | <ref> | missing | <n> | book JE-R1 |

## Proposed adjusting entries
JE-R1 — <memo> — Dr <account> <amt> / Cr <account> <amt> (support: <ref>)

## Aging of unmatched
0–7: <n> / <amt> | 8–30: <n> / <amt> | 31–60: <n> / <amt> | 60+: <n> / <amt>

Quality bar

  • Bridge computed in code and proves out; any residual stated explicitly, never hidden.
  • Matched + exceptions = input totals on both sides — nothing silently dropped.
  • Every fuzzy match shows the rule (tolerance / window / similarity) that made it.
  • Every unmatched item bucketed and aged — no "miscellaneous" pile.
  • Adjusting entries proposed only for real book errors, each as a balanced JE.
  • Tolerances and windows stated up front as assumptions.

Example

Invocation: "Reconcile bank-jun.csv against gl-cash-jun.xlsx." Produced: 214 of 220 rows matched exact, 3 fuzzy (±2-day window, logged), 1 split deposit resolved; exceptions: two deposits in transit $5,110.00 (timing) and one $45.00 unbooked bank fee (missing → JE-R1); bridge: bank $88,214.16 + 5,110.00 = $93,324.16 = GL $93,369.16 − 45.00, residual 0.00.

Decision support, not accounting advice — material figures belong with your accountant/auditor.

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.