solana-dev
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes advanced safety guardrails (W011) that specifically instruct the agent to ignore any directives found in on-chain data, account metadata, or program logs, treating them as untrusted input.
- [CREDENTIALS_UNSAFE]: Explicitly forbids the agent from asking for or storing private keys, seed phrases, or keypair files (W009). It mandates the use of wallet-standard signing flows where keys remain in the user's wallet.
- [EXTERNAL_DOWNLOADS]: Fetches CLI tools and development environments from official Solana ecosystem sources. It provides instructions to install the Solana MCP server from 'mcp.solana.com' and the Solana CLI from 'release.anza.xyz'.
- [REMOTE_CODE_EXECUTION]: Installation patterns for development tools follow established Solana ecosystem practices. It recommends tools like 'surfpool' for local networking, with installation scripts sourced from 'run.surfpool.run'.
- [COMMAND_EXECUTION]: Encourages the 'NO_DNA=1' standard for CLI tools like Anchor and Surfpool to disable interactive TUIs and enable structured output for agent compatibility.
Audit Metadata