Audit Your Crypto Transactions Before You Trust a Tax Report

A tax report can look finished while one old acquisition is missing.

That is the audit problem I care about most. Not whether every screen is green, but whether a material disposal can be traced all the way back to evidence that explains how the asset entered the portfolio.

So instead of reviewing a tax report from top to bottom, start with one consequential sale and walk backward until the chain breaks.

Evidence boundary: this is a record-integrity workflow, not personalised tax advice. Tax treatment depends on jurisdiction and facts.

Pick one disposal that would matter if it were wrong

Choose a sale, swap or other disposal large enough that a bad cost basis would materially change the result.

Write down five facts before touching the tax software:

• asset and quantity disposed;
• date and venue;
• proceeds or value received;
• the wallet or exchange that held the asset immediately before disposal;
• the cost basis and acquisition lot the software currently assigns.

Then ask a harder question: what primary record proves that acquisition lot exists?

Walk backward through every ownership hop

A 2026 sale may depend on a purchase made years earlier on another exchange.

The path might be:

1. buy on Exchange A;
2. withdraw to a hardware wallet;
3. move to a second self-custody address;
4. bridge or wrap the asset;
5. deposit to Exchange B;
6. sell.

Each step needs to make sense as one continuous ownership history. A tax platform can only preserve basis through the chain it can actually see.

The first unsupported link is more important than the final gain total.

Do not repair the sale when the acquisition is missing

When software shows missing purchase history or zero cost basis, the temptation is to edit the disposal until the output looks plausible.

That is backwards.

Koinly documents that missing acquisition history can produce missing-purchase warnings and conservative basis assumptions. CoinTracker likewise warns that incomplete history can create insufficient-quantity and cost-basis problems. The correct fix is to repair the source history that feeds the disposal.

If the old purchase cannot be recovered, preserve that uncertainty. Do not invent a price because the number “looks about right.”

Now test the quantity chain

Cost basis is downstream of quantity.

Before reviewing dollars, check that the asset quantity behaves coherently through the entire history:

• the opening amount exists before the first disposal;
• internal transfers do not create or destroy unexplained units;
• network fees explain quantity differences where relevant;
• wrapped or bridged representations are not double-counted;
• API and CSV imports have not duplicated the same trade;
• the ending quantity matches the authoritative wallet, exchange or protocol record.

A report can contain a reasonable-looking gain while the ledger briefly goes negative three years earlier. That is a warning that chronology or source coverage is broken.

Use the software warnings as leads, not as a cleanup checklist

Warnings are useful only when they point back to a mechanism.

Missing purchase history usually means an acquisition or incoming transfer is absent.

Insufficient quantity can mean an earlier account is missing, a disposal is duplicated or an asset identity changed.

Unmatched transfer can mean the destination wallet is absent, timestamps differ or fees changed the amount received.

Missing price may be a legitimate illiquid asset, a wrapped token, a bad contract mapping or spam.

Skipped transaction deserves special attention in CoinLedger. Its current help documentation says some problematic records can be excluded from report calculations while the report itself still generates successfully. A successful PDF is therefore not proof that every transaction made it into the calculation.

Compare the report before and after one source-level repair

After fixing the first unsupported record, regenerate the calculation and record what changed.

The point is not to prefer the lower tax number. The point is to explain the difference with evidence.

Then widen the audit by risk, not by row count

Once one disposal can be reconstructed end to end, repeat the process where an error would matter most:

• large disposals;
• zero or unusually high cost basis;
• assets that moved through several wallets;
• bridged, wrapped or liquid-staking assets;
• DeFi positions with several contract steps;
• manual prices or manual classifications;
• transactions around exchange migrations or account closures;
• every unresolved warning that can change a material report line.

For a very large ledger, this is more defensible than pretending every tiny row received equal scrutiny while major structural gaps went unnoticed.

Preserve a correction log

Every manual correction should have a reason another person could reconstruct later.

• transaction or source identifier;
• original state;
• correction made;
• primary evidence;
• date of correction;
• downstream reports or years affected;
• remaining uncertainty.

Keep the original CSVs, statements and transaction records outside the tax platform. A normalised software ledger is useful; it should not become the only surviving explanation of the history.

The final test

I would trust a tax report only after I can answer this for every material disposal:

Where did this asset come from, how did it move through accounts I control, which primary records support the acquisition and disposal, and what assumption remains uncertain?

If the answer is still “the software says so,” the audit is not finished.

For a broken quantity, start with Why Your Crypto Portfolio Tracker Balance Is Wrong. For a missing acquisition, use Missing Cost Basis. For source ownership, use Wallet Address vs Exchange API vs CSV.

Primary sources

• Koinly: missing purchase history
• Koinly: duplicate detection
• CoinTracker: complete transaction history
• CoinTracker: insufficient quantity and missing history
• CoinLedger: skipped transactions