The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The installation guide includes a command to fetch and execute the official Rust installer from https://sh.rustup.rs. It also provides instructions to clone the project's source code from its GitHub repository (github.com/QuipNetwork/xq-rs).
[REMOTE_CODE_EXECUTION]: The documentation references the official Rust installation script (https://sh.rustup.rs) which is piped to the shell for execution. As this is a well-known and standard service for the Rust programming language, it is documented here neutrally without escalating the verdict.
[COMMAND_EXECUTION]: The documentation provides examples of typical development commands, including building the software with cargo, running tests, and executing the project's specific CLI tools (xqasm, xqdism, and xqvm). It also includes a Rust code snippet illustrating how to automate these tools using the std::process::Command module.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of external data formats, specifically .xqasm source files and binary bytecode. This ingestion of external content constitutes a potential attack surface for indirect prompt injection if an agent is tasked with analyzing or executing untrusted code. 1. Ingestion points: The xqasm and xqvm CLI utilities, and the StreamReader Rust component. 2. Boundary markers (absent): No explicit instructions to the agent to ignore instructions embedded within the bytecode or source files. 3. Capability inventory: File system access for reading/writing files and sub-process execution via std::process::Command. 4. Sanitization (absent): No evidence of input validation or escaping for the processed quantum optimization models.

aglais-xqvm-quantum-vm