MEV Intelligence

Most tools have clear distinct purposes: building calldata, fetching waves, feeding opportunities, predicting probabilities, recommending builders, and submitting bundles. Some overlap exists among get_liquidation_waves, get_mev_feed, and monitor_crossing_candidates—all list liquidation candidates—but their descriptions differentiate via risk tags, priorities, and cross probability sorting, reducing ambiguity.

All tool names use lowercase_snake_case and follow a verb_noun pattern (get_, build_, monitor_, predict_, recommend_, submit_). While the verb set is not uniform (many get_ but also others), the style is consistent and each verb clearly indicates the action, making the naming predictable and readable.

With 10 tools, the server is well-scoped for the MEV liquidation domain. Each tool covers a distinct step in the workflow—from data discovery to transaction construction to submission—without excessive redundancy or missing essential operations. The count is ideal for an MCP server of this purpose.

The tool surface covers the full liquidation lifecycle: identifying candidates (get_liquidation_waves, get_mev_feed, monitor_crossing_candidates), assessing opportunities (predict_liquidation_probability), building transactions (build_liquidation_tx), choosing a builder (get_builder_recommendation, recommend_bundle_submission), and submitting (submit_bundle). Additional reporting and leaderboard tools provide context. No major gaps that would hinder agent execution are apparent.