The Agent Skills Directory

[SAFE]: The skill focuses on documenting the legitimate use of the SP1 zkVM. All commands and code snippets provided align with the official documentation and best practices for the technology.
[COMMAND_EXECUTION]: Provides instructions for using the cargo prove CLI to manage projects and build RISC-V ELFs. These commands are necessary for the development workflow of the supported technology.
[EXTERNAL_DOWNLOADS]: Mentions sp1up and cargo install for installing the required development tools. These references point to the standard installation methods for the SP1 project.
[REMOTE_CODE_EXECUTION]: Describes the process of compiling Rust source code into executable ELFs and running them within the SP1 SDK's virtual machine environment. This behavior is the primary functionality of a zkVM and is documented for use with the official, reputable toolchain.
[PROMPT_INJECTION]: The skill enables processing external data via sp1_zkvm::io::read and SP1Stdin::write (found in core-program-io.md) for execution within the zkVM. Capability inventory includes cargo prove build and client.execute (found in core-proving-and-verification.md). While explicit boundary markers are not used, the execution is sandboxed within a RISC-V zkVM by design.

sp1