skills/npc-live/clawfirm/solana-skill/Gen Agent Trust Hub

solana-skill

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill manages sensitive Solana private keys and Helius API keys. Private keys are stored encrypted at rest in the user's home directory (~/.config/solana-skill/wallets/) using AES-256-GCM with restricted file permissions (0600/0700).\n- [COMMAND_EXECUTION]: Orchestrates blockchain operations by executing local TypeScript scripts (scripts/send.ts, scripts/swap.ts) via the tsx utility.\n- [PROMPT_INJECTION]: Surface for indirect prompt injection exists due to the processing of external blockchain data.\n
  • Ingestion points: scripts/wallet.ts (getBalance, getFullBalance), scripts/swap.ts (getQuote)\n
  • Boundary markers: Absent\n
  • Capability inventory: Transaction execution (sendSol, sendToken) and token swaps (executeSwap)\n
  • Sanitization: Basic Solana address validation using the @solana/web3.js library
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 06:07 AM
Security Audit — agent-trust-hub — solana-skill