Account Data Matching

Summary #

Use data validation checks to verify that account data matches an expected value. Without appropriate data validation checks, unexpected accounts may be used in an instruction handler.

To implement data validation checks in Rust, simply compare the data stored on an account to an expected value.

if ctx.accounts.user.key() != ctx.accounts.user_data.user {
    return Err(ProgramError::InvalidAccountData.into());
}

In Anchor, you can use a constraint to check whether the given expression evaluates to true. Alternatively, you can use has_one to check that a target account field stored on the account matches the key of an account in the Accounts struct.

Lesson #

Account data matching refers to data validation checks used to verify the data stored on an account matches an expected value. Data validation checks provide a way to include additional constraints to ensure the appropriate accounts are passed into an instruction handler.

This can be useful when accounts required by an instruction handler have dependencies on values stored in other accounts or if an instruction handler is dependent on the data stored in an account.

Missing data validation check #

The example below includes an update_admin instruction handler that updates the admin field stored on an admin_config account.

The instruction handler is missing a data validation check to verify the admin account signing the transaction matches the admin stored on the admin_config account. This means any account signing the transaction and passed into the instruction handler as the admin account can update the admin_config account.

use anchor_lang::prelude::*;
 
declare_id!("Fg6PaFpoGXkYsidMpWTK6W2BeZ7FEfcYkg476zPFsLnS");
 
#[program]
pub mod data_validation {
    use super::*;
    ...
    pub fn update_admin(ctx: Context<UpdateAdmin>) -> Result<()> {
        ctx.accounts.admin_config.admin = ctx.accounts.new_admin.key();
        Ok(())
    }
}
 
#[derive(Accounts)]
pub struct UpdateAdmin<'info> {
    #[account(mut)]
    pub admin_config: Account<'info, AdminConfig>,
    #[account(mut)]
    pub admin: Signer<'info>,
    /// CHECK: This account will not be checked by anchor
    pub new_admin: UncheckedAccount<'info>,
}
 
#[account]
pub struct AdminConfig {
    admin: Pubkey,
}

Add Data Validation Check #

The basic Rust approach to solve this problem is to simply compare the passed in admin key to the admin key stored in the admin_config account, throwing an error if they don't match.

if ctx.accounts.admin.key() != ctx.accounts.admin_config.admin {
    return Err(ProgramError::InvalidAccountData.into());
}

By adding a data validation check, the update_admin instruction handler would only process if the admin signer of the transaction matched the admin stored on the admin_config account.

use anchor_lang::prelude::*;
 
declare_id!("Fg6PaFpoGXkYsidMpWTK6W2BeZ7FEfcYkg476zPFsLnS");
 
#[program]
pub mod data_validation {
    use super::*;
    ...
    pub fn update_admin(ctx: Context<UpdateAdmin>) -> Result<()> {
      if ctx.accounts.admin.key() != ctx.accounts.admin_config.admin {
            return Err(ProgramError::InvalidAccountData.into());
        }
        ctx.accounts.admin_config.admin = ctx.accounts.new_admin.key();
        Ok(())
    }
}
 
#[derive(Accounts)]
pub struct UpdateAdmin<'info> {
    #[account(mut)]
    pub admin_config: Account<'info, AdminConfig>,
    #[account(mut)]
    pub admin: Signer<'info>,
    /// CHECK: This account will not be checked by anchor
    pub new_admin: UncheckedAccount<'info>,
}
 
#[account]
pub struct AdminConfig {
    admin: Pubkey,
}

Use Anchor Constraints #

Anchor simplifies this with the has_one constraint. You can use the has_one constraint to move the data validation check from the instruction handler logic to the UpdateAdmin struct.

In the example below, has_one = admin specifies that the admin account signing the transaction must match the admin field stored on the admin_config account. To use the has_one constraint, the naming convention of the data field on the account must be consistent with the naming on the account validation struct.

use anchor_lang::prelude::*;
 
declare_id!("Fg6PaFpoGXkYsidMpWTK6W2BeZ7FEfcYkg476zPFsLnS");
 
#[program]
pub mod data_validation {
    use super::*;
    ...
    pub fn update_admin(ctx: Context<UpdateAdmin>) -> Result<()> {
        ctx.accounts.admin_config.admin = ctx.accounts.new_admin.key();
        Ok(())
    }
}
 
#[derive(Accounts)]
pub struct UpdateAdmin<'info> {
    #[account(
        mut,
        has_one = admin
    )]
    pub admin_config: Account<'info, AdminConfig>,
    #[account(mut)]
    pub admin: Signer<'info>,
    /// CHECK: This account will not be checked by anchor
    pub new_admin: UncheckedAccount<'info>,
}
 
#[account]
pub struct AdminConfig {
    admin: Pubkey,
}

Alternatively, you can use constraint to manually add an expression that must evaluate to true in order for execution to continue. This is useful when for some reason naming can't be consistent or when you need a more complex expression to fully validate the incoming data.

#[derive(Accounts)]
pub struct UpdateAdmin<'info> {
    #[account(
        mut,
        constraint = admin_config.admin == admin.key()
    )]
    pub admin_config: Account<'info, AdminConfig>,
    #[account(mut)]
    pub admin: Signer<'info>,
    /// CHECK: This account will not be checked by anchor
    pub new_admin: UncheckedAccount<'info>,
}

Lab #

For this lab, we'll create a simple “vault” program similar to the program we used in the Signer Authorization lesson and the Owner Check lesson. Similar to those labs, we'll show in this lab how a missing data validation check could allow the vault to be drained.

1. Starter #

To get started, download the starter code from the starter branch of this repository. The starter code includes a program with two instructions and the boilerplate setup for the test file.

The initialize_vault instruction handler initializes a new Vault account and a new TokenAccount. The Vault account will store the address of a token account, the authority of the vault, and a withdraw destination token account.

The authority of the new token account will be set as the vault, a PDA of the program. This allows the vault account to sign for the transfer of tokens from the token account.

The insecure_withdraw instruction handler transfers all the tokens in the vault account's token account to a withdraw_destination token account.

Info

Notice that this instruction handler does have a signer check for authority and an owner check for vault. However, nowhere in the account validation or instruction handler logic is there code that checks that the authority account passed into the instruction handler matches the authority account on the vault.

use anchor_lang::prelude::*;
use anchor_spl::token::{self, Mint, Token, TokenAccount};
 
declare_id!("J89xWAprDsLAAwcTA6AhrK49UMSAYJJWdXvw4ZQK4suu");
 
pub const DISCRIMINATOR_SIZE: usize = 8;
 
#[program]
pub mod account_data_matching {
    use super::*;
 
    pub fn initialize_vault(ctx: Context<InitializeVault>) -> Result<()> {
        ctx.accounts.vault.token_account = ctx.accounts.token_account.key();
        ctx.accounts.vault.authority = ctx.accounts.authority.key();
        ctx.accounts.vault.withdraw_destination = ctx.accounts.withdraw_destination.key();
        Ok(())
    }
 
    pub fn insecure_withdraw(ctx: Context<InsecureWithdraw>) -> Result<()> {
        let amount = ctx.accounts.token_account.amount;
 
        let seeds = &[b"vault".as_ref(), &[ctx.bumps.vault]];
        let signer = [&seeds[..]];
 
        let cpi_ctx = CpiContext::new_with_signer(
            ctx.accounts.token_program.to_account_info(),
            token::Transfer {
                from: ctx.accounts.token_account.to_account_info(),
                authority: ctx.accounts.vault.to_account_info(),
                to: ctx.accounts.withdraw_destination.to_account_info(),
            },
            &signer,
        );
 
        token::transfer(cpi_ctx, amount)?;
        Ok(())
    }
}
 
#[derive(Accounts)]
pub struct InitializeVault<'info> {
    #[account(
        init,
        payer = authority,
        space = DISCRIMINATOR_SIZE + Vault::INIT_SPACE,
        seeds = [b"vault"],
        bump,
    )]
    pub vault: Account<'info, Vault>,
    #[account(
        init,
        payer = authority,
        token::mint = mint,
        token::authority = vault,
        seeds = [b"token"],
        bump,
    )]
    pub token_account: Account<'info, TokenAccount>,
    pub withdraw_destination: Account<'info, TokenAccount>,
    pub mint: Account<'info, Mint>,
    #[account(mut)]
    pub authority: Signer<'info>,
    pub token_program: Program<'info, Token>,
    pub system_program: Program<'info, System>,
    pub rent: Sysvar<'info, Rent>,
}
 
#[derive(Accounts)]
pub struct InsecureWithdraw<'info> {
    #[account(
        seeds = [b"vault"],
        bump,
    )]
    pub vault: Account<'info, Vault>,
    #[account(
        mut,
        seeds = [b"token"],
        bump,
    )]
    pub token_account: Account<'info, TokenAccount>,
    #[account(mut)]
    pub withdraw_destination: Account<'info, TokenAccount>,
    pub token_program: Program<'info, Token>,
    pub authority: Signer<'info>,
}
 
#[account]
#[derive(Default, InitSpace)]
pub struct Vault {
    token_account: Pubkey,
    authority: Pubkey,
    withdraw_destination: Pubkey,
}

2. Test insecure_withdraw Instruction Handler #

To prove that this is a problem, let's write a test where an account other than the vault's authority tries to withdraw from the vault.

The test file includes the code to invoke the initialize_vault instruction handler using the provider wallet as the authority and then mints 100 tokens to the vault token account.

Add a test to invoke the insecure_withdraw instruction handler. Use fakeWithdrawDestination as the withdrawDestination account and fakeWallet as the authority. Then send the transaction using fakeWallet.

Since there are no checks the verify the authority account passed into the instruction handler matches the values stored on the vault account initialized in the first test, the instruction handler will process successfully and the tokens will be transferred to the fakeWithdrawDestination account.

describe("Account Data Matching", () => {
  ...
  it("allows insecure withdrawal", async () => {
    try {
      const tx = await program.methods
        .insecureWithdraw()
        .accounts({
          vault: vaultPDA,
          tokenAccount: tokenPDA,
          withdrawDestination: fakeWithdrawDestination,
          authority: fakeWallet.publicKey,
        })
        .transaction();
 
      await anchor.web3.sendAndConfirmTransaction(provider.connection, tx, [
        fakeWallet,
      ]);
 
      const tokenAccount = await getAccount(provider.connection, tokenPDA);
      expect(Number(tokenAccount.amount)).to.equal(0);
    } catch (error) {
      throw new Error(
        `Insecure withdraw failed unexpectedly: ${error.message}`,
      );
    }
  });
})

Run anchor test to see that both transactions will complete successfully.

Account Data Matching
    ✔ initializes the vault and mints tokens (879ms)
    ✔ allows insecure withdrawal (431ms)

3. Add secure_withdraw Instruction Handler #

Let's go implement a secure version of this instruction handler called secure_withdraw.

This instruction handler will be identical to the insecure_withdraw instruction handler, except we'll use the has_one constraint in the account validation struct (SecureWithdraw) to check that the authority account passed into the instruction handler matches the authority account on the vault account. That way only the correct authority account can withdraw the vault's tokens.

use anchor_lang::prelude::*;
use anchor_spl::token::{self, Mint, Token, TokenAccount};
 
declare_id!("Fg6PaFpoGXkYsidMpWTK6W2BeZ7FEfcYkg476zPFsLnS");
 
pub const DISCRIMINATOR_SIZE: usize = 8;
 
#[program]
pub mod account_data_matching {
    use super::*;
    ...
    pub fn secure_withdraw(ctx: Context<SecureWithdraw>) -> Result<()> {
        let amount = ctx.accounts.token_account.amount;
 
        let seeds = &[b"vault".as_ref(), &[ctx.bumps.vault]];
        let signer = [&seeds[..]];
 
        let cpi_ctx = CpiContext::new_with_signer(
            ctx.accounts.token_program.to_account_info(),
            token::Transfer {
                from: ctx.accounts.token_account.to_account_info(),
                authority: ctx.accounts.vault.to_account_info(),
                to: ctx.accounts.withdraw_destination.to_account_info(),
            },
            &signer,
        );
 
        token::transfer(cpi_ctx, amount)?;
        Ok(())
    }
}
 
#[derive(Accounts)]
pub struct SecureWithdraw<'info> {
    #[account(
        seeds = [b"vault"],
        bump,
        has_one = token_account,
        has_one = authority,
        has_one = withdraw_destination,
    )]
    pub vault: Account<'info, Vault>,
    #[account(
        mut,
        seeds = [b"token"],
        bump,
    )]
    pub token_account: Account<'info, TokenAccount>,
    #[account(mut)]
    pub withdraw_destination: Account<'info, TokenAccount>,
    pub token_program: Program<'info, Token>,
    pub authority: Signer<'info>,
}

4. Test secure_withdraw Instruction Handler #

Now let's test the secure_withdraw instruction handler with two tests: one that uses fakeWallet as the authority and one that uses wallet as the authority. We expect the first invocation to return an error and the second to succeed.

describe("account-data-matching", () => {
  ...
  it("prevents unauthorized secure withdrawal", async () => {
    try {
      const tx = await program.methods
        .secureWithdraw()
        .accounts({
          vault: vaultPDA,
          tokenAccount: tokenPDA,
          withdrawDestination: fakeWithdrawDestination,
          authority: fakeWallet.publicKey,
        })
        .transaction();
 
      await anchor.web3.sendAndConfirmTransaction(provider.connection, tx, [
        fakeWallet,
      ]);
 
      throw new Error("Secure withdraw should have failed but didn't");
    } catch (error) {
      expect(error).to.be.an("error");
      console.log("Expected error occurred:", error.message);
    }
  });
 
  it("allows secure withdrawal by authorized user", async () => {
    try {
      await new Promise((resolve) => setTimeout(resolve, 1000));
 
      await mintTo(
        provider.connection,
        wallet.payer,
        mint,
        tokenPDA,
        wallet.payer,
        100,
      );
 
      await program.methods
        .secureWithdraw()
        .accounts({
          vault: vaultPDA,
          tokenAccount: tokenPDA,
          withdrawDestination,
          authority: wallet.publicKey,
        })
        .rpc();
 
      const tokenAccount = await getAccount(provider.connection, tokenPDA);
      expect(Number(tokenAccount.amount)).to.equal(0);
    } catch (error) {
      throw new Error(`Secure withdraw failed unexpectedly: ${error.message}`);
    }
  });
})

Run anchor test to see that the transaction using an incorrect authority account will now return an Anchor Error while the transaction using the correct accounts complete successfully.

"Program J89xWAprDsLAAwcTA6AhrK49UMSAYJJWdXvw4ZQK4suu invoke [1]",
"Program log: Instruction: SecureWithdraw",
"Program log: AnchorError caused by account: vault. Error Code: ConstraintHasOne. Error Number: 2001. Error Message: A has one constraint was violated.",
"Program log: Left:",
"Program log: GprrWv9r8BMxQiWea9MrbCyK7ig7Mj8CcseEbJhDDZXM",
"Program log: Right:",
"Program log: 2jTDDwaPzbpG2oFnnqtuHJpiS9k9dDVqzzfA2ofcqfFS",
"Program J89xWAprDsLAAwcTA6AhrK49UMSAYJJWdXvw4ZQK4suu consumed 11790 of 200000 compute units",
"Program J89xWAprDsLAAwcTA6AhrK49UMSAYJJWdXvw4ZQK4suu failed: custom program error: 0x7d1"

Note that Anchor specifies in the logs the account that causes the error (AnchorError caused by account: vault).

✔ prevents unauthorized secure withdrawal
✔ allows secure withdrawal by authorized user (1713ms)

And just like that, you've closed up the security loophole. The theme across most of these potential exploits is that they're quite simple. However, as your programs grow in scope and complexity, it becomes increasingly easy to miss possible exploits. It's great to get in a habit of writing tests that send instructions that shouldn't work. The more the better. That way you catch problems before you deploy.

If you want to take a look at the final solution code you can find it on the solution branch of the repository.

Challenge #

Just as with other lessons in this unit, your opportunity to practice avoiding this security exploit lies in auditing your own or other programs.

Take some time to review at least one program and ensure that proper data checks are in place to avoid security exploits.

Remember, if you find a bug or exploit in somebody else's program, please alert them! If you find one in your own program, be sure to patch it right away.

Completed the lab?

Push your code to GitHub and tell us what you thought of this lesson!