The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of various CLI tools including forge, cast, and anvil for interacting with the Ethereum blockchain and managing local test nodes. It provides examples for creating transactions, calling contracts, and managing local network state.
[DYNAMIC_EXECUTION]: In references/testing.md, the skill explains the use of the vm.ffi (Foreign Function Interface) cheatcode. This feature allows the execution of arbitrary shell commands during the testing process. While a powerful capability, it is a standard and documented feature of the Foundry framework typically used for differential testing.
[EXTERNAL_DOWNLOADS]: The skill frontmatter specifies foundry as a dependency to be installed via brew. Additionally, references/dependencies.md details how to manage project dependencies using forge install (Git submodules) and soldeer (a Solidity package manager). These are established practices within the smart contract development ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates an environment that ingests external configuration files (foundry.toml) and source code (*.sol). An attacker-controlled file could potentially influence the behavior of the underlying CLI tools or the AI agent during development. However, this is an inherent characteristic of development-focused skills, and the documentation emphasizes best practices for security and testing.
Ingestion points: Project configuration in foundry.toml and Solidity contract files in src/, test/, and script/ directories.
Boundary markers: None identified; the skill assumes a trusted local development context.
Capability inventory: The skill uses subprocess calls to execute forge, cast, and anvil commands, and supports the high-privilege vm.ffi capability for shell execution.
Sanitization: Standard for technical documentation; no specific sanitization of input files is described.

foundry-solidity