pump-ai-agents
Warn
Audited by Socket on Jul 9, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill’s general crypto/Pump SDK purpose is coherent, but it goes beyond passive agent guidance by having the agent install and run an unverified MCP server via unpinned `npx -y`. Because the external CLI provenance was not clearly verified and such tooling may receive wallet or RPC credentials and enable financial actions, the install-trust and scope risks are disproportionate to a documentation-style skill.
Confidence: 85%Severity: 82%
Audit Metadata