injective-evm-developer
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code files (Solidity, TypeScript, JavaScript), which constitutes an attack surface for indirect prompt injection.\n
- Ingestion points: SKILL.md specifies activation on .sol, hardhat.config., foundry.toml, *.ts, and *.js files.\n
- Capability inventory: The agent utilizes developer tools (Hardhat, Foundry) and MCP servers to interact with these files.\n
- Boundary markers: The instructions do not include specific delimiters or warnings to the agent regarding embedded instructions in these files.\n
- Sanitization: No sanitization or validation of the ingested code content is performed.\n- [EXTERNAL_DOWNLOADS]: The skill references external assets and tools from the vendor's infrastructure and third-party sources.\n
- Fetches Solidity interfaces from the official InjectiveLabs GitHub repository.\n
- References official Injective documentation and faucet services.\n
- Recommends the installation of a third-party skill: whackur/solidity-agent-toolkit.\n- [COMMAND_EXECUTION]: Provides shell commands for project setup, contract deployment, and verification.\n
- Includes commands for npx skills add, npx hardhat verify, and deployment script execution.\n- [CREDENTIALS_UNSAFE]: The provided Hardhat configuration template correctly utilizes environment variables (process.env.PRIVATE_KEY) for secret management, ensuring that private keys are not hardcoded.
Audit Metadata