phantom-wallet-mcp

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to download the @phantom/mcp-server package from the NPM registry. This represents standard integration with the official Phantom wallet infrastructure.\n- [REMOTE_CODE_EXECUTION]: The skill executes code from the downloaded @phantom/mcp-server package at runtime to perform wallet operations.\n- [COMMAND_EXECUTION]: The environment setup requires the execution of a shell command via npx to launch the MCP server.\n- [DATA_EXFILTRATION]: The skill interacts with sensitive local session data stored at ~/.phantom-mcp/session.json, which is used for persistent authentication between sessions.\n- [INDIRECT_PROMPT_INJECTION]: \n
  • Ingestion points: The buy_token tool fetches Solana swap quotes from the external Phantom API, processing dynamic data at runtime.\n
  • Boundary markers: The instructions do not define specific delimiters or clear warnings for the agent to ignore instructions embedded in API responses.\n
  • Capability inventory: The skill possesses high-impact capabilities including signing transactions, transferring tokens, and signing messages across multiple blockchains.\n
  • Sanitization: There is no explicit evidence of sanitization or validation of the contents returned by external API endpoints before they are processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 03:04 PM