arcium
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The documentation includes a command to fetch and execute an installation script from the vendor's official domain (
install.arcium.com). This is documented as a standard setup procedure for the developer toolchain. - [COMMAND_EXECUTION]: Instructions involve using the
arciumCLI anddockerto manage local development environments, build circuits, and run tests. These are essential for the primary purpose of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation and processing of Arcis circuit and Solana program code, which are surfaces for handling external user data.
- Ingestion points: Processes user-provided Rust logic in circuit and program files (e.g.,
lib.rs). - Boundary markers: Utilizes Arcium's MPC boundaries (Shared/Mxe encryption) and Anchor's account context to define access and reveal capabilities.
- Capability inventory: Involves local command execution, blockchain state modification, and communication with Arx compute nodes via RPC.
- Sanitization: Relies on Anchor's validation macros and Arcium's cryptographic secret-sharing protocols to ensure secure data handling and authorized transitions.
Audit Metadata