index-bot

Index Bot

Build a weighted basket of assets from one instruction, buy the whole basket in one pass, and rebalance on a cadence when weights drift. This skill is a thin strategy layer: it decides what to buy and how much, then delegates every quote, approval, swap, and signing step to existing skills.

Runtime Compatibility: This skill uses AskUserQuestion for interactive prompts. If AskUserQuestion is not available in your runtime, collect the same parameters through natural language conversation instead.

Overview

From a single prompt (for example "equal-weight basket of the top 5 RWAs, rebalance weekly"), index-bot:

1. Parses the basket spec (explicit assets with weights, or a top-N ranking request).
2. Resolves each asset to a token address and computes per-leg sizing.
3. Applies guardrails (spend cap, token allowlist, dry-run).
4. Delegates each buy to the swap-integration Trading API flow on the target chain.
5. Records target weights and the last rebalance in a state file.

On a later invocation by the host scheduler, it reads current positions, computes drift versus target weights, generates the adjusting swaps, and delegates them.

The trigger is twofold: a prompt to create the basket, and the host scheduler (cron or the agent runtime wake-up) for each rebalance run.

Prerequisites

This skill does not reimplement swap execution. It depends on:

• swap-integration (uniswap-trading): the only execution path. Every buy and rebalance leg goes through its Trading API flow (check_approval then quote then swap, then sign and broadcast). Do not reimplement any of it.
• v4-sdk-integration (uniswap-trading): LP execution, only if a leg also seeds or manages a liquidity position. Default index baskets are spot-only and do not need this.
• viem-integration (uniswap-viem): accounts, signing, transaction broadcast, and reading on-chain balances for the drift calculation.

Read these plugin references before acting and treat them as ground truth:

• selected target-chain template: chainId, chain name, contract addresses, tradable token source, funding constraints, market-data availability, and template-specific caveats. For the reference Robinhood Chain template, see ../../references/robinhood-chain.md.
• ../../references/execution-model.md: the Trading API requirement, execution modes, restrictions, and disclaimer rules.
• ../../references/strategy-state.md: the shared state file and scheduler pattern.

Template inputs

The selected target-chain template must provide:

• chain id, chain name, native gas token, and RPC / read path.
• deployed Uniswap router, Permit2, and quoter / state-reader addresses needed by delegated execution and valuation.
• optional v4 PoolManager / PositionManager addresses if the basket also manages LP positions.
• tradable token resolution rules, including any token list, ranking source, or allowlist source.
• funding constraints, market-data availability, transfer-restriction caveats, and market-hours guidance.

Workflow

Step 1: Parse the basket spec

Extract the basket definition from the prompt:

If the prompt asks for a top-N ranking ("top 5 RWAs"), there is no default ranking source, so do NOT invent or guess a ranking. If the selected template provides an indexer or ranking source and the operator instructs you to use it, use that source. Otherwise ask the user for an explicit asset list via AskUserQuestion and proceed from that list. If any other required parameter is missing, including per-run or per-period spend caps, use AskUserQuestion with structured options to collect it.

Step 2: Resolve tokens and per-leg sizing

For each asset:

1. Resolve the symbol to a token address from the selected target-chain template and its token source. Do not maintain a local registry. If a requested asset cannot resolve from the template's token source, stop and report it rather than guessing.
2. Compute the target notional per leg: legNotional = totalSize * weight. Price any cross-token sizing with Uniswap quotes (Trading API /quote via swap-integration, or V4Quoter via v4-sdk-integration), never an external feed or venue-specific API.
3. Record the funding-token amount to spend on that leg.

Verify each resolved address is a contract before sizing. If an asset cannot be resolved on the target chain, stop and report it rather than guessing.

Step 3: Guardrails

Before any execution:

• Ask the operator for any missing per-run or per-period spend cap using AskUserQuestion (or natural language if unavailable); include the funding token or denomination used for comparison. If either cap remains unset, do not enter autonomous mode.
• Confirm the basket total is within the configured spend cap (per run and per period).
• Confirm every leg token is on the operator's allowlist.
• Run a dry-run first that prints each planned leg (token, amount, weight) for review.
• Confirm equity market hours if any leg is a RWA (off-hours liquidity may be thin).
• Per leg, check the wallet can cover that leg: it must hold enough of the funding token (tokenIn) for the leg's buy amount and enough native gas to broadcast. If a leg is short on either, skip that leg and report it, then continue the rest of the basket. Follow the selected template's funding constraints; the skill does not auto-fund, bridge, or top up unless that behavior is explicitly provided outside the skill.

Step 4: Buy the basket (delegate)

For each leg, delegate to the swap-integration Trading API flow: check_approval then quote then swap, then sign and broadcast via viem. Pass the target chain id and that chain's router / token addresses from the selected template.

Do NOT reimplement quoting, approvals, swap-body construction, or signing. Sequence legs with a short delay to respect Trading API rate limits, and handle a failed leg without abandoning the rest of the basket (report it and continue).

Step 5: Record target weights

After the buys succeed, write the target weights and the rebalance metadata to the state file (see ../../references/strategy-state.md). This is what later drift calculations compare against.

Rebalance loop

On each scheduled invocation:

1. Read state to load target weights and lastRebalanceAt.
2. Check idempotency: if this cadence period was already rebalanced, skip.
3. Read current on-chain balances for each leg via viem and value them in the funding token using Uniswap quotes (Trading API /quote via swap-integration, or V4Quoter / StateView via v4-sdk-integration) to get current weights. Do not use venue-specific APIs or external price feeds; see ../../references/execution-model.md (Data and pricing).
4. Compute drift per leg: drift = currentWeight - targetWeight. Compare the absolute drift |drift| against the threshold so both overweight (drift > 0) and underweight (drift < 0) legs are caught. If no leg's |drift| exceeds the drift threshold, do nothing and update lastRebalanceAt.
5. Generate the adjusting swaps: sell overweight legs, buy underweight legs, to return each leg to its target.
6. Delegate each adjusting swap to the swap-integration Trading API flow exactly as in Step 4.
7. Update state with the new lastRebalanceAt and any changed positions.

Execution mode

This skill exposes the shared execution mode from ../../references/execution-model.md:

Default to confirm. Autonomous mode requires all of: a spend cap (per run and per period), a token allowlist, a dry-run first, and a kill switch. A basket buy is many transactions, so under confirm summarize the full basket and confirm once per run where the runtime allows, otherwise per leg.

State

State follows ../../references/strategy-state.md. index-bot stores the target weights and the last rebalance so drift can be computed on the next run:

Each run reads state first and updates it only after a successful broadcast, so reruns within a period do not double-buy or double-rebalance.

Restrictions and Disclaimers

RWA gating is template-specific and may be enforced at the token level (transfer-restricted ERC-20s), so a leg can revert at transfer time even when the router accepts the quote. This skill must:

1. Handle transfer-restriction reverts per leg gracefully and report which leg reverted.
2. Not assume a pool-level or router-level allowlist exists.
3. Respect equity market hours; some RWAs may have off-hours liquidity limits.
4. Surface the financial disclaimers in the repo root DISCLAIMER.md before executing.

See ../../references/execution-model.md for the full restrictions and disclaimer rules.

Input validation

Before interpolating ANY user-provided value into generated code, API calls, or shell commands:

• Token addresses: MUST match ^0x[a-fA-F0-9]{40}$; reject otherwise.
• Amounts and notionals: MUST be non-negative numeric values matching ^[0-9]+\.?[0-9]*$.
• Weights: MUST each be in [0, 1] and the full set MUST sum to 1 (within a small tolerance). Reject a basket whose weights do not sum to 1.
• Chain id: MUST be the operator-configured target chain id, read from the selected template rather than hardcoded.
• API keys: MUST come from environment variables, never hardcoded.
• REJECT any input containing shell metacharacters: ;, |, &, $, `, (, ), >, <, \, ', ", newlines.