Glossary OS

Plain meaning

Start with the shortest useful explanation before going deeper.

A vulnerability where a program accepts an arbitrary program account from the caller and invokes it via Cross-Program Invocation (CPI) without verifying it matches a known, trusted program ID, effectively letting an attacker substitute a malicious program that executes under the victim program's authority or manipulates accounts the victim program passes to it. A common pattern is accepting a token_program account without checking it equals spl_token::ID, so the attacker passes a lookalike program that records or drains account data. Prevention requires hard-coding or explicitly checking the program ID before every CPI call.

Mental model

Use the quick analogy first so the term is easier to reason about when you meet it in code, docs, or prompts.

Think of it as a building block that connects one definition to the larger Solana system around it.

Technical context

Place the term inside its Solana layer so the definition is easier to reason about.

Failure modes, audits, attack surfaces, and safe design patterns.

Why builders care

Turn the term from vocabulary into something operational for product and engineering work.

This term unlocks adjacent concepts quickly, so it works best when you treat it as a junction instead of an isolated definition.

AI handoff

Use this compact block when you want to give an agent or assistant grounded context without dumping the entire page.

Arbitrary CPI (arbitrary-cpi)
Category: Security
Definition: A vulnerability where a program accepts an arbitrary program account from the caller and invokes it via Cross-Program Invocation (CPI) without verifying it matches a known, trusted program ID, effectively letting an attacker substitute a malicious program that executes under the victim program's authority or manipulates accounts the victim program passes to it. A common pattern is accepting a token_program account without checking it equals spl_token::ID, so the attacker passes a lookalike program that records or drains account data. Prevention requires hard-coding or explicitly checking the program ID before every CPI call.
Related: Cross-Program Invocation (CPI), Program

Glossary Copilot

Ask grounded Solana questions without leaving the glossary.

Use glossary context, relationships, mental models, and builder paths to get structured answers instead of generic chat output.

Open full Copilot workspace

Question

Explain this code

Optional: paste Anchor, Solana, or Rust code so the Copilot can map primitives back to glossary terms.

Ask a glossary-grounded question

The Copilot will answer using the current term, related concepts, mental models, and the surrounding glossary graph.

Concept graph

See the term as part of a network, not a dead-end definition.

These branches show which concepts this term touches directly and what sits one layer beyond them.

Branch

Cross-Program Invocation (CPI)

A mechanism for one program to call another program's instructions during execution. CPIs enable composability—e.g., a DeFi program can call the Token Program to transfer tokens. CPI depth is limited to 4 levels. The caller passes accounts and instruction data, and the callee runs with the same transaction context.

Instruction invoke()

Open term

Branch

Program

Executable code deployed on-chain, equivalent to a smart contract on other blockchains. Programs are stateless—they store no data themselves but read/write data in separate accounts they own. Programs are compiled to SBF bytecode and loaded via the BPF Loader. Every program has a unique Program ID (its account's public key).

Account BPF (Berkeley Packet Filter)

Open term

Next concepts to explore

Keep the learning chain moving instead of stopping at one definition.

These are the next concepts worth opening if you want this term to make more sense inside a real Solana workflow.

Programming Model

Cross-Program Invocation (CPI)

A mechanism for one program to call another program's instructions during execution. CPIs enable composability—e.g., a DeFi program can call the Token Program to transfer tokens. CPI depth is limited to 4 levels. The caller passes accounts and instruction data, and the callee runs with the same transaction context.

Open term

Programming Model

Program

Executable code deployed on-chain, equivalent to a smart contract on other blockchains. Programs are stateless—they store no data themselves but read/write data in separate accounts they own. Programs are compiled to SBF bytecode and loaded via the BPF Loader. Every program has a unique Program ID (its account's public key).

Open term

Security

Audit Severity Levels

A standardized classification system for security audit findings: Critical (immediate fund loss or protocol compromise, must fix before deployment), High (significant risk of loss under specific conditions), Medium (limited risk or requires unlikely preconditions), Low (minor issues with minimal impact), and Informational (code quality, gas optimization, or best practice suggestions). Severity ratings help teams prioritize remediation and communicate risk to users. Most Solana audit firms (OtterSec, Ackee, sec3) follow this five-tier scale.

Open term

Security

Account Revival Attack

An exploit that resurrects an account that a program has logically closed within the same transaction by sending lamports back to it before the transaction finalizes, causing its on-chain data — which was never securely wiped — to re-appear as a funded, seemingly valid account in future transactions. Because the Solana runtime keeps an account alive as long as it holds any lamports, transferring even 1 lamport back to a closed-but-not-wiped account prevents its deletion and allows an attacker to reuse its stale state. The defense is to explicitly overwrite account data with a closed discriminator and to use force-defund patterns so any lamports transferred in during the same transaction are immediately drained.

Open term

Related terms

Follow the concepts that give this term its actual context.

Glossary entries become useful when they are connected. These links are the shortest path to adjacent ideas.

Programming Modelcpi

Cross-Program Invocation (CPI)

A mechanism for one program to call another program's instructions during execution. CPIs enable composability—e.g., a DeFi program can call the Token Program to transfer tokens. CPI depth is limited to 4 levels. The caller passes accounts and instruction data, and the callee runs with the same transaction context.

Open term

Programming Modelprogram

Program

Executable code deployed on-chain, equivalent to a smart contract on other blockchains. Programs are stateless—they store no data themselves but read/write data in separate accounts they own. Programs are compiled to SBF bytecode and loaded via the BPF Loader. Every program has a unique Program ID (its account's public key).

Open term

Stay in the same layer and keep building context.

These entries live beside the current term and help the page feel like part of a larger knowledge graph instead of a dead end.

Security

Missing Signer Check

A vulnerability where a program accepts an account in a privileged role (e.g., admin, authority, payer) without verifying that the account actually signed the transaction, allowing any caller to impersonate that authority by simply passing the target pubkey as an instruction account. In native Solana programs, the check requires asserting account.is_signer == true; in Anchor, the Signer<'info> type enforces this automatically. Exploitation lets an attacker bypass all access control gated on authority equality checks, making it one of the most critical and commonly audited vulnerabilities in Solana programs.

Open term

Security

Missing Owner Check

A vulnerability where a program deserializes and trusts account data without first confirming that the account is owned by the expected program, allowing an attacker to substitute a maliciously crafted account owned by a different program whose byte layout happens to satisfy the deserialization. On Solana, every account stores a 32-byte owner field set to the program that created it; native programs must assert account.owner == &expected_program_id, while Anchor's Account<'info, T> wrapper performs this check automatically. Failure to validate ownership can lead to complete auth bypass if an attacker can construct a fake account whose data parses into a struct with elevated privileges.

Open term

Security

PDA Substitution Attack

A vulnerability where a program derives a PDA internally but accepts an externally supplied account as that PDA without re-deriving and comparing the address, allowing an attacker to pass a different PDA (derived from attacker-controlled seeds) that the program will treat as legitimate. Because PDAs are deterministic, the only way to guarantee account identity is to call Pubkey::find_program_address (or equivalent) with the expected seeds inside the program and assert the result equals the supplied key. Anchor's seeds and bump constraints on the Account type automate this re-derivation and equality check.

Open term

Security

Account Reloading Attack

A vulnerability in Anchor programs where a program reads an account's data before making a CPI call, the CPI modifies that account's lamports or data, but the program continues using the stale pre-CPI snapshot instead of reloading the account from the runtime. In Anchor, after a CPI the account reference still holds the pre-call data unless account.reload() is explicitly called, meaning balance checks, state assertions, or further computation can operate on incorrect values. Attackers can exploit this to pass checks using an initial account state that the CPI subsequently invalidates.

Open term