agentleFS
agiprolabsagiprolabs/claude-trading-skills/skills/dex-execution/SKILL.md

dex-execution

Skill366 starsChanged 8 months agoMITView on GitHub

Solana DEX swap execution via Jupiter aggregator including quoting, transaction building, signing, and confirmation

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every swap follows this seven-step pipeline.

    line 18
  2. Distinctive:

    Always show these fields before proceeding:

    line 37
  3. Very specific:

    Use "auto" for most cases.

    line 170
  4. Very specific:

    asLegacyTransaction: Use legacy format instead of versioned transactions.

    line 175
  5. Distinctive:

    Final check: verify token balance changed as expected

    line 204

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

DEX Execution — Solana Swap Execution via Jupiter

Execute token swaps on Solana through Jupiter, the dominant DEX aggregator routing across Raydium, Orca, Meteora, Phoenix, Lifinity, and 20+ other venues.

Overview

Jupiter aggregates liquidity across all major Solana DEXes to find optimal swap routes. A single swap may split across multiple pools and hop through intermediate tokens to minimize price impact. The Jupiter v6 API handles route discovery, transaction building, and fee optimization — your code handles quoting, user confirmation, signing, and submission.

Base URL: https://quote-api.jup.ag/v6

Execution Pipeline

Every swap follows this seven-step pipeline. Never skip steps 2-3 (display and confirm).

Step 1 — Get Quote

import httpx

params = {
    "inputMint": "So11111111111111111111111111111111111111112",   # SOL
    "outputMint": "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v", # USDC
    "amount": 1_000_000_000,  # 1 SOL in lamports
    "slippageBps": 50,        # 0.5%
}
resp = httpx.get("https://quote-api.jup.ag/v6/quote", params=params)
quote = resp.json()

Step 2 — Display Quote to User

Always show these fields before proceeding:

FieldSource
Input amountquote["inAmount"] (in token decimals)
Output amountquote["outAmount"]
Minimum receivedquote["otherAmountThreshold"]
Price impactquote["priceImpactPct"]
Routequote["routePlan"] — DEXes used
SlippageThe slippageBps you requested

Step 3 — Require User Confirmation

⚠️  SWAP PREVIEW
    Selling:   1.000 SOL
    Buying:    ~142.50 USDC
    Min recv:  141.79 USDC (0.5% slippage)
    Impact:    0.01%
    Route:     Raydium V4 → USDC

    Proceed? [y/N]

NEVER proceed without explicit "yes" from the user.

Step 4 — Build Transaction

swap_body = {
    "quoteResponse": quote,
    "userPublicKey": "YourPubkeyBase58...",
    "wrapAndUnwrapSol": True,
    "dynamicComputeUnitLimit": True,
    "prioritizationFeeLamports": "auto",
}
resp = httpx.post("https://quote-api.jup.ag/v6/swap", json=swap_body)
swap_data = resp.json()
swap_tx = swap_data["swapTransaction"]  # base64-encoded transaction

Step 5 — Sign Transaction

import base64
from solders.transaction import VersionedTransaction
from solders.keypair import Keypair

raw_tx = base64.b64decode(swap_tx)
tx = VersionedTransaction.from_bytes(raw_tx)
keypair = Keypair.from_base58_string(os.environ["WALLET_PRIVATE_KEY"])
tx.sign([keypair])

Step 6 — Submit Transaction

rpc_url = os.environ.get("SOLANA_RPC_URL", "https://api.mainnet-beta.solana.com")
signed_bytes = bytes(tx)
payload = {
    "jsonrpc": "2.0", "id": 1,
    "method": "sendTransaction",
    "params": [
        base64.b64encode(signed_bytes).decode(),
        {"encoding": "base64", "skipPreflight": False,
         "maxRetries": 3, "preflightCommitment": "confirmed"}
    ],
}
resp = httpx.post(rpc_url, json=payload)
sig = resp.json()["result"]
print(f"Submitted: https://solscan.io/tx/{sig}")

Step 7 — Confirm Transaction

import time

for attempt in range(30):
    payload = {
        "jsonrpc": "2.0", "id": 1,
        "method": "getSignatureStatuses",
        "params": [[sig], {"searchTransactionHistory": False}],
    }
    resp = httpx.post(rpc_url, json=payload)
    status = resp.json()["result"]["value"][0]
    if status and status.get("confirmationStatus") in ("confirmed", "finalized"):
        print(f"Confirmed at slot {status['slot']}")
        break
    time.sleep(2)
else:
    print("Transaction not confirmed within 60s — check explorer")

Jupiter API v6 Endpoints

EndpointMethodPurpose
/quoteGETGet best-price quote with routing
/swapPOSTBuild a swap transaction from a quote
/swap-instructionsPOSTGet individual instructions (advanced)
/price?ids=token1,token2GETSimple price lookup (v2)
/tokensGETList all supported tokens

See references/jupiter_api.md for full parameter and response documentation.

Key Parameters

Slippage (slippageBps)

Token TypeRecommended RangeNotes
SOL, USDC, major tokens50-100 (0.5-1%)Stable liquidity
Mid-cap tokens100-300 (1-3%)Variable liquidity
PumpFun / meme tokens500-2000 (5-20%)Thin books, high volatility
New launches (<1h old)1000-3000 (10-30%)Extreme volatility

Dynamic Slippage

Set dynamicSlippage: true in the swap request to let Jupiter auto-adjust slippage based on current market conditions. Preferred for most use cases.

Priority Fees (prioritizationFeeLamports)

Priority fees determine transaction ordering within a block.

LevelmicroLamportsWhen to Use
Low10,000-50,000Normal conditions
Medium50,000-200,000Moderate congestion
High200,000-1,000,000High congestion / time-sensitive
Urgent1,000,000-5,000,000Meme coin launches, NFT mints
Auto"auto"Jupiter estimates for you

Use "auto" for most cases. For fine-grained control, query Helius getPriorityFeeEstimate.

Other Parameters

  • onlyDirectRoutes: Skip multi-hop routing. Faster but may get worse price.
  • asLegacyTransaction: Use legacy format instead of versioned transactions. Required for some older wallets.
  • maxAccounts: Limit accounts in transaction (default 64). Lower values reduce route options but improve confirmation reliability.
  • platformFeeBps: Integrator fee in basis points. Taken from output amount.
  • wrapAndUnwrapSol: Auto wrap/unwrap SOL ↔ wSOL (default true).
  • dynamicComputeUnitLimit: Auto-set compute budget based on simulation.

Priority Fee Estimation

# Using Helius getPriorityFeeEstimate
helius_url = f"https://mainnet.helius-rpc.com/?api-key={HELIUS_KEY}"
payload = {
    "jsonrpc": "2.0", "id": 1,
    "method": "getPriorityFeeEstimate",
    "params": [{"accountKeys": [input_mint, output_mint],
                "options": {"recommended": True}}],
}
resp = httpx.post(helius_url, json=payload)
fee = resp.json()["result"]["priorityFeeEstimate"]

Transaction Confirmation Strategy

See references/transaction_lifecycle.md for the full Solana transaction lifecycle.

  1. Submit with skipPreflight: False to catch obvious errors
  2. Poll getSignatureStatuses every 2 seconds for up to 60 seconds
  3. If not confirmed: rebuild transaction with fresh blockhash and retry (max 3 attempts)
  4. Blockhash expiry: transactions expire ~60 seconds after blockhash was fetched
  5. Final check: verify token balance changed as expected

Error Handling

ErrorCauseRecovery
"Slippage tolerance exceeded"Price moved beyond slippageBpsIncrease slippage or retry
"Insufficient funds"Not enough input token or SOL for feesCheck balance before quoting
"Transaction expired"Blockhash too oldRebuild with fresh blockhash
"Transaction simulation failed"Various — check logsParse simulation logs for root cause
"Too many accounts"Route uses too many accountsSet maxAccounts lower or use onlyDirectRoutes
HTTP 429Rate limitedBack off and retry with exponential delay
HTTP 400 "No route found"No liquidity path existsCheck token mints are correct, try larger slippage

Safety Requirements

These are non-negotiable requirements for any execution code.

  1. ALWAYS show quote details (amounts, price impact, route) before execution
  2. ALWAYS require explicit user confirmation — never auto-execute
  3. Default to simulation mode — do not sign or submit unless explicitly enabled
  4. Never store or log private keys — load from env vars, use immediately, discard
  5. Never use 100% slippage — this is a common scam/exploit vector
  6. Verify token addresses — confirm mints match expected tokens before swapping
  7. Check price impact — warn if >2%, block if >10% unless user overrides
  8. Maintain SOL reserve — keep 0.05 SOL minimum for rent and future fees

See references/safety_checklist.md for the complete pre/during/post execution checklist.

Integration with Other Skills

SkillIntegration
slippage-modelingEstimate optimal slippageBps based on token liquidity profile
liquidity-analysisVerify pool depth supports trade size before quoting
position-sizingCalculate trade amount based on risk parameters
risk-managementEnforce portfolio-level exposure limits before execution
jupiter-apiUnderlying API documentation for Jupiter endpoints
helius-apiPriority fee estimation and transaction monitoring

Files

References

  • references/jupiter_api.md — Complete Jupiter v6 API parameter and response reference
  • references/transaction_lifecycle.md — Solana transaction lifecycle, priority fees, retry strategies
  • references/safety_checklist.md — Pre/during/post execution verification checklist

Scripts

  • scripts/get_quote.py — Fetch and display Jupiter swap quotes with route analysis
  • scripts/simulate_swap.py — Build and simulate swap transactions without submitting

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.