# Research And Trade > Research a token and execute a trade if it passes due diligence. Autonomous research-to-trade pipeline: researches the token, evaluates risk, and only trades if the risk assessment approves. Stops and reports if risk is too high. Use when user wants "research X and buy if it looks good" or "due diligence then trade." - Skill: `modbender/research-and-trade` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add modbender/research-and-trade` - Raw SKILL.md: https://api.skillmd.com/api/skills/modbender/research-and-trade/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: DevOps & Infra - Author: modbender (https://skillmd.com/u/modbender) - Updated: 2026-09-09 - Page: https://skillmd.com/skills/modbender/research-and-trade --- # Research and Trade ## Overview This is the autonomous research-to-execution pipeline. Instead of manually calling four different agents and wiring their outputs together, this skill runs the full expert workflow in one command: research a token, find the best pool, assess risk, and -- only if the risk assessment approves -- execute the trade. **Why this is 10x better than calling agents individually:** 1. **Compound context**: Each agent receives the accumulated findings from all prior agents. The risk-assessor doesn't just evaluate a swap in isolation -- it sees the token-analyst's liquidity warnings, the pool-researcher's depth analysis, and the exact trade size, enabling a far richer risk assessment than standalone invocation. 2. **Automatic risk gating**: A VETO at any stage short-circuits the pipeline immediately. No wasted gas, no wasted time, and you get a full explanation of why. 3. **Single command for a 4-step expert workflow**: Manually coordinating token research, pool selection, risk evaluation, and trade execution takes significant time and expertise. This compresses it into one natural-language request. 4. **Progressive disclosure**: You see each stage's findings as they complete, not just a final result. If the pipeline stops at risk assessment, you still get the full research report. ## When to Use Activate when the user says anything like: - "Research UNI and buy if it looks good" - "Due diligence on AAVE then trade" - "Investigate and trade ARB" - "Should I buy LINK? If so, do it" - "Is PEPE safe to trade? Buy $500 worth if yes" - "Research X, assess risk, and swap if it passes" - "Autonomous trade: research then execute" - "Check out TOKEN and buy some if the risk is acceptable" **Do NOT use** when the user just wants research without trading (use `research-token` instead) or just wants to execute a swap without research (use `execute-swap` instead). ## Parameters | Parameter | Required | Default | How to Extract | | ------------- | -------- | ---------- | ------------------------------------------------------------------ | | token | Yes | -- | Token to research and potentially buy: "UNI", "AAVE", or 0x addr | | amount | Yes | -- | Trade size: "$500", "1 ETH worth", "0.5 ETH" | | chain | No | ethereum | Target chain: "ethereum", "base", "arbitrum" | | riskTolerance | No | moderate | "conservative", "moderate", "aggressive" | | action | No | buy | "buy" (swap into token) or "sell" (swap out of token) | | payWith | No | WETH | Token to spend: "WETH", "USDC", etc. | If the user doesn't provide an amount, **ask for it** -- never guess a trade size. ## Workflow ``` RESEARCH-AND-TRADE PIPELINE ┌─────────────────────────────────────────────────────────────────────┐ │ │ │ Step 1: RESEARCH (token-analyst) │ │ ├── Token metadata, liquidity, volume, risk factors │ │ ├── Cross-chain presence │ │ └── Output: Token Research Report │ │ │ │ │ ▼ (feeds into Step 2) │ │ │ │ Step 2: POOL ANALYSIS (pool-researcher) │ │ ├── Find all pools for {token}/{payWith} on {chain} │ │ ├── Rank by fee APY, depth, utilization │ │ ├── Analyze depth at trade size (can it handle $X?) │ │ └── Output: Pool Research Report + Best Pool Selection │ │ │ │ │ ▼ (feeds into Step 3 with COMPOUND CONTEXT) │ │ │ │ Step 3: RISK ASSESSMENT (risk-assessor) │ │ ├── Receives: token risks + pool risks + trade size + slippage │ │ ├── Evaluates: slippage, liquidity, smart contract, token risk │ │ ├── Decision: APPROVE / CONDITIONAL_APPROVE / VETO / HARD_VETO │ │ └── Output: Risk Assessment Report │ │ │ │ │ ▼ CONDITIONAL GATE │ │ ┌───────────────────────────────────────────────────┐ │ │ │ APPROVE → Proceed to Step 4 │ │ │ │ COND. APPROVE → Show conditions, ask user │ │ │ │ VETO → STOP. Show research + reason. │ │ │ │ HARD VETO → STOP. Non-negotiable. │ │ │ └───────────────────────────────────────────────────┘ │ │ │ (only if APPROVE or user confirms CONDITIONAL) │ │ ▼ │ │ │ │ Step 4: USER CONFIRMATION │ │ ├── Present: research summary + risk score + swap quote │ │ ├── Ask: "Proceed with this trade?" │ │ └── User must explicitly confirm │ │ │ │ │ ▼ │ │ │ │ Step 5: EXECUTE (trade-executor) │ │ ├── Execute swap through safety-guardian pipeline │ │ ├── Monitor transaction confirmation │ │ └── Output: Trade Execution Report │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` ### Step 1: Research (token-analyst) Delegate to `Task(subagent_type:token-analyst)` with: - Token symbol or address - Target chain - Request: full due diligence report **What to pass to the agent:** ``` Research this token for a potential trade: - Token: {token} - Chain: {chain} - Trade size: {amount} Provide a full due diligence report: liquidity across all pools, volume profile (24h/7d/30d), risk factors, and a trading recommendation with maximum trade size at < 1% price impact. ``` **Present to user after completion:** ```text Step 1/5: Token Research Complete Token: UNI (Uniswap) on Ethereum Total Liquidity: $85M across 24 pools 24h Volume: $15M | 7d Volume: $95M Volume Trend: Stable Risk Factors: None significant Max Trade (< 1% impact): $2.5M Proceeding to pool analysis... ``` **Gate check:** If the token-analyst reports critical risk factors (total liquidity < $100K, no pools found, token not verified), present findings and ask the user if they want to continue before proceeding. ### Step 2: Pool Analysis (pool-researcher) Delegate to `Task(subagent_type:pool-researcher)` with the token research output: ``` Find the best pool for trading {token}/{payWith} on {chain}. Context from token research: - Total liquidity: {from Step 1} - Dominant pool: {from Step 1} - Risk factors: {from Step 1} Trade details: - Trade size: {amount} - Direction: {action} (buying/selling {token}) Analyze all pools for this pair across fee tiers. For each pool, report: fee APY, TVL, liquidity depth at the trade size, and price impact estimate. Recommend the best pool for this specific trade. ``` **Present to user after completion:** ```text Step 2/5: Pool Analysis Complete Best Pool: WETH/UNI 0.3% (V3, Ethereum) Pool TVL: $42M Price Impact: ~0.3% for your trade size Fee Tier: 0.3% (3000 bps) Proceeding to risk assessment... ``` ### Step 3: Risk Assessment (risk-assessor) Delegate to `Task(subagent_type:risk-assessor)` with **compound context** from Steps 1 and 2: ``` Evaluate risk for this proposed swap: Operation: swap {amount} {payWith} for {token} Pool: {best pool from Step 2} Chain: {chain} Risk tolerance: {riskTolerance} Token research context (from token-analyst): {Full token research summary from Step 1} Pool analysis context (from pool-researcher): {Full pool analysis from Step 2} Evaluate all applicable risk dimensions: slippage, liquidity, smart contract risk. Provide a clear APPROVE / CONDITIONAL_APPROVE / VETO / HARD_VETO decision. ``` **Conditional gate logic after risk-assessor returns:** | Decision | Action | | -------------------- | ------------------------------------------------------------------------------------ | | **APPROVE** | Present risk summary, proceed to Step 4 (user confirmation) | | **CONDITIONAL_APPROVE** | Show conditions (e.g., "split into 2 tranches"). Ask user: "Accept conditions?" | | **VETO** | **STOP.** Show full research report + risk assessment + veto reason. Suggest alternatives. | | **HARD_VETO** | **STOP.** Show reason. Non-negotiable -- do not offer to proceed. | **Present to user (APPROVE case):** ```text Step 3/5: Risk Assessment Complete Decision: APPROVE Composite Risk: LOW Slippage Risk: LOW (0.3% estimated) Liquidity Risk: LOW (pool TVL 840x trade size) Smart Contract Risk: LOW (V3, 18-month-old pool) Ready for your confirmation... ``` **Present to user (VETO case):** ```text Step 3/5: Risk Assessment -- VETOED Decision: VETO Reason: Price impact of 4.2% exceeds moderate risk tolerance (max 2%) Research Summary: Token: SMALLCAP ($180K total liquidity) Best Pool: WETH/SMALLCAP 1% (V3, $95K TVL) Your trade size ($5,000) represents 5.3% of pool TVL Suggestions: - Reduce trade size to < $1,000 for acceptable slippage - Use a limit order instead: "Submit limit order for SMALLCAP" - Try a different chain if more liquidity exists elsewhere Pipeline stopped. No trade executed. ``` ### Step 4: User Confirmation Before executing any trade, present a clear summary and ask for explicit confirmation: ```text Trade Confirmation Required Research: UNI — $85M liquidity, stable volume, no risk factors Risk: APPROVED (LOW composite risk) Swap Details: Sell: 0.5 WETH (~$980) Buy: ~28.5 UNI Pool: WETH/UNI 0.3% (V3, Ethereum) Impact: ~0.3% Gas: ~$8 estimated Proceed with this trade? (yes/no) ``` **Only proceed to Step 5 if the user explicitly confirms.** ### Step 5: Execute (trade-executor) Delegate to `Task(subagent_type:trade-executor)` with the full pipeline context: ``` Execute this swap: - Sell: {amount} {payWith} - Buy: {token} - Pool: {best pool address from Step 2} - Chain: {chain} - Slippage tolerance: {derived from risk assessment} - Risk assessment: APPROVED, composite risk {level} The token has been researched (liquidity: {X}, volume: {Y}) and risk-assessed (slippage: {Z}, liquidity: {W}). Proceed with execution through the safety pipeline. ``` **Present final result:** ```text Step 5/5: Trade Executed Sold: 0.5 WETH ($980.00) Received: 28.72 UNI ($985.50) Pool: WETH/UNI 0.3% (V3, Ethereum) Slippage: 0.28% (within tolerance) Gas: $7.20 Tx: https://etherscan.io/tx/0x... ────────────────────────────────────── Pipeline Summary ────────────────────────────────────── Research: UNI — $85M liquidity, stable, no risk flags Pool: WETH/UNI 0.3% — best depth for trade size Risk: APPROVED (LOW) Execution: Success — 28.72 UNI received Total cost: $987.20 (trade + gas) ``` ## Output Format ### Successful Pipeline (all 5 steps) ```text Research and Trade Complete Token: {symbol} ({name}) on {chain} Research: {1-line summary from token-analyst} Pool: {pool pair} {fee}% ({version}, {chain}) Risk: {decision} ({composite_risk}) Trade: Sold: {amount} {payWith} (${usd_value}) Received: {amount} {token} (${usd_value}) Impact: {slippage}% Gas: ${gas_cost} Tx: {explorer_link} Pipeline: Research -> Pool -> Risk -> Confirm -> Execute (all passed) ``` ### Vetoed Pipeline (stopped at risk) ```text Research and Trade -- Risk Vetoed Token: {symbol} ({name}) on {chain} Research: {1-line summary} Pool: {best pool found} Risk: VETOED — {reason} Details: {risk dimension scores} Suggestions: - {mitigation 1} - {mitigation 2} Pipeline: Research -> Pool -> Risk (VETOED) -- No trade executed. ``` ## Important Notes - **This skill always researches first.** It never skips to trading. If the user just wants a quick swap without research, redirect them to `execute-swap`. - **Risk gating is non-negotiable for HARD_VETO.** If the risk-assessor issues a HARD_VETO (unverified token, pool TVL < $1K, price impact > 10%), the pipeline stops. The user cannot override this. - **VETO is informational.** For a regular VETO, present the full research and explain why. The user can then choose to use `execute-swap` directly if they want to proceed at their own risk -- but this skill will not do it. - **Compound context is the key differentiator.** The risk-assessor is dramatically more useful when it has the token-analyst's risk factors and the pool-researcher's depth analysis, compared to calling it standalone with just a swap request. - **Progressive output keeps the user informed.** Don't wait until the end to show results. After each agent completes, show a brief summary so the user knows what's happening. - **Amount is required.** Never assume a trade size. If the user says "research and buy UNI" without an amount, ask: "How much would you like to trade?" ## Error Handling | Error | User-Facing Message | Suggested Action | | ----------------------------- | ------------------------------------------------------------------------ | ----------------------------------------- | | Token not found | "Could not find token {X} on {chain}." | Check spelling or provide contract address| | No pools found | "No Uniswap pools found for {token}/{payWith} on {chain}." | Try different pay token or chain | | Token-analyst fails | "Token research failed: {reason}. Cannot proceed without due diligence." | Try again or use research-token directly | | Pool-researcher fails | "Pool analysis failed. Research completed but cannot find optimal pool." | Try execute-swap with manual pool choice | | Risk-assessor VETO | "Risk assessment vetoed this trade: {reason}." | Reduce amount, try different token/pool | | Risk-assessor HARD_VETO | "Trade blocked: {reason}. This cannot be overridden." | The trade is unsafe at any size | | Trade-executor fails | "Trade execution failed: {reason}. Research and risk data preserved." | Check wallet, balance, gas; retry | | Safety check fails | "Safety limits exceeded. Check spending limits with check-safety." | Wait for limit reset or adjust limits | | User declines confirmation | "Trade cancelled. Research and risk data are shown above for reference." | No action needed | | Wallet not configured | "No wallet configured. Cannot execute trades." | Set up wallet with setup-agent-wallet | | Insufficient balance | "Insufficient {payWith} balance: have {X}, need {Y}." | Reduce amount or acquire more tokens |