The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (LOW): The skill references and provides instructions for installing external Python packages halmos and certora-cli for formal verification. While these are legitimate developer tools, they are retrieved from external repositories at runtime.
Evidence: Installation commands pip install halmos and pip install certora-cli found in references/formal-verification.md.
[PROMPT_INJECTION] (LOW): The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and process smart contract code and test files to generate scripts and tests. A malicious codebase could potentially contain hidden instructions aimed at influencing the agent's logic or the generated deployment scripts.
Ingestion points: The skill interacts with files in tests/, scripts/, and src/ directories.
Boundary markers: No explicit boundary markers or 'ignore' instructions for external data are used in the provided templates.
Capability inventory: The skill utilizes shell commands for forge, just, pip, halmos, and certoraRun, including on-chain transaction broadcasts.
Sanitization: No evidence of input sanitization for code or data read from the local repository.

web3-foundry