agentleFS
iktok90-designiktok90-design/ai-smart-contract-auditor/skills/fee-on-transfer/SKILL.md

fee-on-transfer

Skill36 starsChanged 5 months agoMITView on GitHub

Detect fee-on-transfer / deflationary / rebasing token accounting bugs — crediting the *passed amount* instead of the measured balance delta. Activate whenever code calls transfer/transferFrom and then credits, mints shares for, or records the literal amount argument, in deposits, AMM swaps, lending collateral, vaults, staking, or bridges — without measuring balanceAfter - balanceBefore.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read-only quoting that never credits balances is unaffected.

    line 66

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Fee-on-transfer token accounting detection

When this applies

Trigger on any of:

  • Deposit/stake/addLiquidity that records the amount argument after transferFrom
  • Vault/ERC-4626 deposit minting shares proportional to the requested assets, not received
  • AMM/router updating reserves with the input amount rather than actual receipt
  • Lending protocols crediting collateral by passed amount
  • Bridges locking/burning amount then minting the same amount on the far side
  • Any flow assuming received == amount for arbitrary ERC-20s

Detection patterns

Credit passed amount, not delta (HIGH)

function deposit(uint256 amount) external {
    token.transferFrom(msg.sender, address(this), amount);
    balanceOf[msg.sender] += amount;          // ← amount, but FoT token delivered less
    totalDeposited        += amount;
}

Signal: the contract believes it holds more than it does. On a fee-on-transfer token (USDT with fee switch enabled, PAXG, many meme tokens), totalDeposited exceeds the real balance — last withdrawers are insolvent, or an attacker mints excess shares/LP.

AMM reserve desync (HIGH)

Router records input amountIn into reserves but the pair received amountIn - fee. k invariant is computed on phantom reserves, mispricing the swap and letting an attacker drain the shortfall over time (the Uniswap-V2-style "supports FoT" *SupportingFeeOnTransferTokens functions exist precisely because of this).

Rebasing double-count (MEDIUM / HIGH)

uint256 shares = amount * totalShares / totalAssets;   // totalAssets from stored var

Signal: stored totalAssets vs live token.balanceOf(this) diverge as a positive-rebasing token (stETH, aTokens, OHM) grows. Crediting against the stale stored figure under- or over-mints shares; an attacker times deposits around rebase.

Bridge mint mismatch (CRITICAL)

Lock amount on chain A, emit message, mint amount on chain B — but only amount - fee was locked. Repeated deposits accumulate an unbacked mint surplus that can be withdrawn on A by honest users, draining the lockbox.

Severity rubric

PatternSeverityNotes
Bridge mints unbacked surplusCriticalCross-chain insolvency
Vault/LP shares minted on phantom amountHighDirect over-mint, fund loss
AMM reserve recorded as passed amountHighMispricing, drainable shortfall
Lending collateral over-creditedHighUnder-collateralized borrow
Rebasing stored-vs-live totalAssets driftMediumTiming-dependent
FoT token on a hard whitelist of non-FoT tokensInfoSurface closed by policy

Remediation patterns

  1. Measure the delta: uint256 before = token.balanceOf(address(this)); token.transferFrom(...); uint256 received = token.balanceOf(address(this)) - before; and credit received.
  2. Reject FoT on entry — if received != amount, revert (OpenZeppelin SafeERC20 does not do this for you).
  3. Use live balance, not stored totals, for rebasing tokens, or wrap them (wstETH) so balances are non-rebasing.
  4. Token allowlist for assets known not to take a transfer fee; document the assumption.
  5. On bridges, lock the measured received amount and mint exactly that.

False-positive notes

  • A protocol that only ever handles a fixed, audited, non-FoT token (e.g. a WETH-only vault) is safe — note the assumption rather than flagging High.
  • *SupportingFeeOnTransferTokens router variants already measure deltas — don't flag.
  • Read-only quoting that never credits balances is unaffected.

Related

  • [[token-compatibility]] — broader weird-ERC20 behaviors (no return value, blocklists)
  • [[erc4626-inflation]] — share-mint math interacts with miscounted assets
  • [[cross-chain-messaging]] — bridge mint/lock mismatch surface

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.