pump-ai-agents

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill documents standard agent integration patterns for the Solana ecosystem, providing templates for instruction files and skill registries.
  • [COMMAND_EXECUTION]: Defines terminal management rules (isBackground, auto-kill) and MCP server configuration using npx with vendor-affiliated packages (@pump-fun/mcp-server). These are standard operational configurations for the intended platform.
  • [CREDENTIALS_UNSAFE]: The skill explicitly mandates security hardening for sensitive data, including instructions to use 0600 file permissions for keypairs and zeroing out key material after use to prevent exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 03:13 AM
Security Audit — agent-trust-hub — pump-ai-agents