whalehunter-strategy

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

• Third-party content exposure detected (high risk: 0.72). The runtime LLM context is fed by the whalehunter_long_entry / whalehunter_short_entry action’s `context:
• type: signalwhere{{signal_whalehunter_signals}}comes from theexternal_scanner (whalehunter_signals) that ingests Senpi Discover–sourced trader data (outsider-authored free-form fields like username/reasons`), creating an indirect prompt-injection path into the LLM prompt.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly a trading agent: it manages two funded wallets, requires a user-scoped SENPI_AUTH_TOKEN, and references concrete trading operations and order semantics (create_position, close_position, edit_position, cancel_order, ratchet_stop_, strategy_close, FEE_OPTIMIZED_LIMIT orders, leverage caps, per-position margin, maxSlots, etc.). These are specific market-order / wallet management capabilities (crypto/trading execution), not generic tooling. The doc even prescribes which execution calls must or must not be invoked, indicating the skill's primary purpose is placing and managing financial positions. Therefore it grants direct financial execution authority.