The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents core operations such as tronbox compile, migrate, and test which execute local code for development. As described in the migration documentation, these scripts are executed in a VM context to manage deployment and testing states.\n- [EXTERNAL_DOWNLOADS]: The tronbox unbox command is documented as a primary feature used to download project templates from remote repositories. The documentation references well-known and trusted sources such as the official TRON Protocol GitHub organization.\n- [CREDENTIALS_UNSAFE]: While the framework handles sensitive blockchain private keys, the documentation promotes high-security hygiene by instructing users to use environment variables and .env files. Example configurations provided use obvious dummy placeholders to prevent accidental exposure.\n- [PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it facilitates the processing of user-provided Solidity source code and JavaScript migration scripts.\n
Ingestion points: Contract source files and scripts located in the contracts/ and migrations/ directories as defined in the configuration.\n
Boundary markers: Not explicitly specified; the tool is designed to execute valid code directly.\n
Capability inventory: File system writes for build artifacts, script execution during migrations/testing, and network access to blockchain nodes via tronWeb and ethers.\n
Sanitization: Includes the use of @solidity-parser/parser to handle dependency resolution during contract flattening.

tronbox