Reference

LegibilityReport

Every simulate call returns a LegibilityReport. Its JSON shape is stable across patch versions, safe to pipe into jq, render in a UI, or persist in a database for later review.

Top-level shape

LegibilityReport.json{
  "tx_signature_preview": "4dLPchZ8d2eCX2eobPtT9eN4DHw...",
  "fee_payer":            "4kfEfEk7HrCLpdqo3vtrMYYF9e...",
  "uses_durable_nonce":   true,
  "durable_nonce_warning": "Transaction uses a durable nonce ...",
  "instructions":   [ DecodedInstruction, ... ],
  "account_diffs":  [ AccountDiff, ... ],
  "token_transfers": [ TokenTransfer, ... ],
  "overall_risk":   "Critical",
  "human_summary":  [ "[System Program] ...", ... ],
  "simulation_logs": [ "Program 11111... invoke [1]", ... ],
  "simulation_success": true,
  "simulation_error":   null
}

field	type	description
tx_signature_preview	string	The first 88-char signature of the tx, or '(unsigned)' if the tx has no signatures. Used as a human-readable identifier in logs and UI.
fee_payer	string	Base58 pubkey of the first static account key in the message, which is always the fee payer by Solana convention.
uses_durable_nonce	boolean	True iff the first instruction is a System Program AdvanceNonceAccount. This is the only way the runtime permits durable-nonce transactions, so the detection is exact.
durable_nonce_warning	string \| null	Human-readable warning emitted when uses_durable_nonce is true. Null otherwise.
instructions	DecodedInstruction[]	One entry per top-level instruction in the message, in on-wire order. Never empty for a well-formed transaction.
account_diffs	AccountDiff[]	Per-account state diffs from the simulation pre/post snapshots. Empty in offline mode or when no writable account changed.
token_transfers	TokenTransfer[]	Token transfer events synthesized from SPL Token / Token-2022 account layout in the pre/post diffs. Does not require a top-level Transfer instruction — CPI-only transfers are still captured.
overall_risk	RiskLevel	The merged verdict. One of: Low, Medium, High, Critical. See /docs/risk-model for the exact escalation rules.
human_summary	string[]	Ordered human-readable lines suitable for printing verbatim in a terminal or rendering as a list in a UI. Always non-empty.
simulation_logs	string[]	The logs field from simulateTransaction, or ['(offline mode — simulation skipped)'] in offline mode.
simulation_success	boolean	True if the simulation returned no error. Always true in offline mode.
simulation_error	string \| null	The TransactionError formatted as a string, or null if the simulation succeeded.

DecodedInstruction

field	type	description
program_id	string	Base58 pubkey of the program the instruction targets.
program_name	string	Friendly name the decoder assigned: System Program, SPL Token, Token-2022, Squads v4, Jupiter v6, Drift v2, Kamino Lend, MarginFi v2, or Unknown.
instruction_name	string	Specific instruction the decoder matched: Transfer, AdvanceNonceAccount, config_transaction_execute, etc. For Unknown programs, this is also 'Unknown'.
summary	string	A one-sentence description of the instruction's effect, templated with account pubkeys where the decoder can resolve them.
accounts	string[]	The base58 pubkeys of every account referenced by this instruction's accounts index list, in order.
risk	RiskLevel	The per-instruction risk level. The classifier takes the max of every instruction's risk as the starting point for the overall verdict.
risk_reasons	string[]	Human-readable reasons the decoder attached to this risk level. May be empty for LOW-risk instructions.

AccountDiff

field	type	description
address	string	Base58 pubkey of the account that changed.
owner_before	string \| null	Owning program of the account before the simulation.
owner_after	string \| null	Owning program after the simulation. A change here is escalated to CRITICAL by the classifier.
lamports_before	number	Lamport balance before the simulation.
lamports_after	number	Lamport balance after the simulation.
lamports_delta	number	Signed delta (lamports_after - lamports_before). Large negative deltas escalate the overall verdict.
data_changed	boolean	Whether any byte of the account's data region changed.
data_len_before	number	Length of the data region before the simulation.
data_len_after	number	Length of the data region after the simulation.

TokenTransfer

field	type	description
mint	string	Base58 pubkey of the mint the transfer belongs to.
from	string	Source token account, or '?' if the source could not be determined from the diff alone.
to	string	Destination token account, or '?' if not determinable.
ui_amount	number	Raw amount as a floating-point value. Decimals are not yet applied — multiply client-side.
raw_amount	number	Raw amount in the mint's base units.
decimals	number	Mint decimals. Always zero in v0.1.0 — fetch from the mint account client-side if you need proper UI formatting.

RiskLevel

A four-variant enum serialized as one of the following exact strings:

"Low" | "Medium" | "High" | "Critical"