Knuckles-Team polymarket-liquidity-mm Workflow
Dual-sided market making with automated inventory skew balancing and capital recycling via position merging
Step 0: mcp_market_data
Monitor target market orderbook depth and calculate best bid/ask spreads Expected: bid_ask_depth
Step 1: mcp_portfolio [depends_on: Step 0]
Analyze active token inventory and portfolio balance ratio Expected: inventory_weight
Step 2: mcp_risk [depends_on: Step 1]
Compute inventory skew quoting offsets (lower ask if long, higher bid if short) Expected: quoting_price_offsets
Step 3: place-market-making-orders [skill: mcp_orders] [depends_on: Step 2]
Place dual-sided limit orders on target market to capture spread Expected: dual_limit_orders
Step 4: refresh-market-making-orders [skill: mcp_orders] [depends_on: Step 3]
Cancel stale or out-of-range quotes and dynamically adjust offsets based on volatility Expected: updated_market_making_state
Execution
Run this workflow as a dependency-ordered DAG. Steps with no unmet depends_on run in parallel; dependents run after their prerequisites complete.
- Run first (in parallel): Step 0 — mcp_market_data
- After level 0: Step 1 — mcp_portfolio
- After level 1: Step 2 — mcp_risk
- After level 2: Step 3 — place-market-making-orders
- After level 3: Step 4 — refresh-market-making-orders
Execution: If graph-os is reachable, offload the whole DAG via graph_orchestrate action=execute_workflow (or the kg-delegate skill) for true parallel/swarm execution. Otherwise execute the steps natively in dependency order: run steps with no unmet depends_on in parallel, then their dependents.