Learn-skills.dev contract-preflight
Dry-run Stacks contract calls against mainnet state before broadcasting — catches errors, prevents wasted gas
git clone https://github.com/NeverSight/learn-skills.dev
T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/skills-md/aibtcdev/skills/contract-preflight" ~/.claude/skills/neversight-learn-skills-dev-contract-preflight && rm -rf "$T"
data/skills-md/aibtcdev/skills/contract-preflight/SKILL.mdContract Pre-Flight
Dry-run any Stacks contract call against live mainnet state without broadcasting. Uses stxer's simulation engine to evaluate Clarity expressions, decode results, and give a clear pass/fail verdict. If the simulation returns
(err ...) or a runtime error, the skill blocks broadcast and explains why.
What it does
Before you broadcast a contract call, this skill creates a simulation session, runs your Clarity expression against the current chain state, and tells you whether it would succeed or fail. No gas spent. No on-chain state changed. Just a verdict: safe to broadcast, or not.
Why agents need it
On-chain transaction failures cost gas and abort visibly. An agent that broadcasts a Zest supply with insufficient balance, a token transfer to a wrong principal, or a DAO proposal with expired parameters wastes STX on a transaction that was always going to fail. This skill eliminates that category of error entirely.
Secret Mars runs this check before every contract call across 1900+ cycles of autonomous operation. Zero aborted transactions since adopting the pattern.
Commands
doctor
doctorPre-flight checks: stxer API reachability, simulation session creation, runtime detection.
bun run contract-preflight/contract-preflight.ts doctor
run --action=simulate
run --action=simulateSimulate a single contract call. Returns decoded Clarity result and broadcast recommendation.
bun run contract-preflight/contract-preflight.ts run \ --action=simulate \ --sender SP4DXVEC16FS6QR7RBKGWZYJKTXPC81W49W0ATJE \ --contract SP2VCQJGH7PHP2DJK7Z0V48AGBHQAW3R3ZW1QF4N.zsbtc-v2-0 \ --expression '(contract-call? .zsbtc-v2-0 get-balance tx-sender)'
run --action=batch
run --action=batchSimulate a sequence of contract calls in a single session. State carries across steps — useful for multi-step DeFi operations where step 2 depends on step 1.
bun run contract-preflight/contract-preflight.ts run \ --action=batch \ --sender SP4DXVEC16FS6QR7RBKGWZYJKTXPC81W49W0ATJE \ --steps '[ {"contract":"SP2VCQJGH7PHP2DJK7Z0V48AGBHQAW3R3ZW1QF4N.zsbtc-v2-0","expression":"(contract-call? .zsbtc-v2-0 get-balance tx-sender)"}, {"contract":"SM3VDXK3WZZSA84XXFKAFAF15NNZX32CTSG82JFQ4.sbtc-token","expression":"(contract-call? .sbtc-token get-balance tx-sender)"} ]'
install-packs
install-packsNo additional packages required. Uses
fetch() and commander.
Safety notes
- Read-only by design. This skill never broadcasts. It simulates only.
- Session isolation. Each simulation runs in a fresh stxer session. No cross-contamination between runs.
- Expression length limit. Capped at 2,000 characters to prevent abuse.
- Step limit. Maximum 20 steps per batch to prevent runaway simulations.
- Timeout. 15-second timeout per API call. Fails cleanly on timeout.
- No private keys needed. Simulation uses sender address, not signing keys.
- Honest verdict. If the Clarity expression returns
, the skill reports it as unsafe. No sugar-coating.(err ...)
Output contract
Simulate (single call)
{ "status": "success | error", "action": "simulate", "data": { "session_id": "d1c27b645459c702feae3a7a637a4777", "result": { "outcome": "ok", "decoded": "(ok uint 276016)", "raw_hex": "070100000000000000000000000000043630", "safe_to_broadcast": true }, "recommendation": "Simulation passed. Safe to broadcast this contract call." }, "error": null }
Batch (multi-step)
{ "status": "success | error", "action": "batch", "data": { "session_id": "abc123", "total_steps": 2, "passed": 2, "failed": 0, "results": [ {"step": 0, "outcome": "ok", "decoded": "(ok uint 276016)", "safe": true}, {"step": 1, "outcome": "ok", "decoded": "(ok uint 204206)", "safe": true} ], "recommendation": "All steps passed. Safe to broadcast the transaction sequence." }, "error": null }
Simulation Proof
These simulations were run on mainnet state (2026-04-10) without broadcasting:
| Test | Session | Expression | Result | Verdict |
|---|---|---|---|---|
| Balance read | d1c27b645459c702feae3a7a637a4777 | on zsbtc-v2-0 | | Safe |
| Impossible transfer | (second session) | on sbtc-token | | Blocked |
The impossible transfer would have aborted on-chain and wasted gas. The simulation caught it for free.
Architecture
[Agent wants to call contract] ↓ [contract-preflight simulate] ↓ [Create stxer session] → [Eval Clarity expression against mainnet state] ↓ ↓ [Ok result] [Err result] ↓ ↓ "Safe to broadcast" "DO NOT broadcast"
Use cases
- DeFi operations: Verify Zest supply/withdraw, Bitflow swaps, ALEX trades before broadcast.
- Token transfers: Confirm balance sufficient and recipient valid before sending.
- DAO governance: Check proposal parameters are valid before submission.
- Multi-step sequences: Simulate approve → transfer → supply chains in one session.
- Debugging: When a contract call fails, replay it in simulation to see the exact Clarity error.
Limitations
- Simulates against current block state. Cannot predict future state changes (MEV, concurrent txs).
- Stxer API is a third-party service. If it's down, the skill reports the outage but cannot simulate.
- Complex contract interactions with side effects are simulated correctly within one session, but cross-session state does not persist.
- Does not validate post-conditions — only the Clarity return value.
Origin
Winner of AIBTC x Bitflow Skills Pay the Bills competition. Original author: @secret-mars Competition PR: https://github.com/BitflowFinance/bff-skills/pull/258