vc-ai-infra-scout

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow ingests outsider-authored free text from Twitter/X and LinkedIn via SELAT-native advanced_search (fundraising news and investor-thesis tweets) and Tavily POST /search (LinkedIn fundraising posts), which are populated by public user submissions.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly performs paid calls and payment settlements: it states the agent "pays for the data" across named payment rails (x402 via Circle Gateway, SELAT Router, x402 on Base), compiles each step into selat-pay calls, includes a per-run maxAmount and per-step caps, and documents selat-pay probe and --pay (to settle real payments). These are specific payment gateways/rails and an execution CLI (selat-pay/Circle Gateway/Base) — i.e., explicit financial execution capability.