Security audit
Security checks for vulnerabilities and agentic risk
This skill is purpose-aligned for BaseAlpha onboarding and token launch, but it asks an unpinned runtime package to handle a live private key for irreversible blockchain actions.
Review this before installing. Use only a dedicated low-value wallet, pin or verify the MCP package before giving it BASEALPHA_PRIVATE_KEY, expect real Base transactions that may spend gas, and confirm that the BaseAlpha/Supabase endpoints are the endpoints you intend to trust.
The skill instructs users to execute an unpinned package via npx @basealpha/mcp, which fetches the latest published code at runtime. In a workflow that also supplies a live private key through environment variables, a compromised upstream package, typosquat, or malicious update could immediately exfiltrate secrets or submit unauthorized on-chain transactions.
The one-shot onboarding flow describes apply → claim → prepare → sign → send → register but does not prominently warn that this will perform real blockchain transactions using the provided private key. In this context, omission is dangerous because operators may treat the skill as routine setup rather than a live token launch with irreversible financial effects.
The raw HTTP section instructs users to submit wallet identity data, signed claim material, and launch-related payloads to external Supabase endpoints without a clear warning about data transmission and trust boundaries. In a security-sensitive onboarding and launch flow, that omission can lead users to expose sensitive metadata or signed artifacts without understanding they are being sent off-host to third-party infrastructure.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
# 1) Apply (no API key)
curl -X POST "$API/v1/agents/apply" \
-H 'Content-Type: application/json' \
-d '{"wallet":"0x…","agentName":"fleet-bot","runtimeTag":"openclaw","metadataUrl":"https://…","description":"…"}'