agentleFS
sunshineluyaosunshineluyao/Moss-Mcp-Transaction/skills/monad-safe-transfer-preview/SKILL.md

Monad Testnet Safe Transfer Preview

Skill1 starsChanged 3 months agoMITView on GitHub

Preview a native MON transfer on Monad Testnet before it reaches a wallet. Fetches live on-chain data (balance, gas estimate, network state) and evaluates nine safety rules. Returns READY_FOR_WALLET_REVIEW only when all checks pass; any warning yields BLOCKED. No signing, no broadcasting.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    Before any network call, capture the user's exact intent: sender address, recipient address, and transfer amount as a decimal string.

    line 39
  2. Very specific:

    All token amounts MUST be represented as decimal strings (e.g.

    line 51
  3. Very specific:

    If any preflight check yields a warning — chain ID mismatch, RPC timeout, invalid address, zero address, non-positive amount, insufficient …

    line 81

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Monad Testnet Safe Transfer Preview Skill

This skill implements a preview-only safety layer for native MON transfers on the Monad Testnet (chain ID 10143). It is inspired by the safety model described in the Moss documentation, adapted for Monad Testnet where Moss itself is mainnet-only.

Safety Rules

RECORD_INTENT

Before any network call, capture the user's exact intent: sender address, recipient address, and transfer amount as a decimal string. Store these values immutably for the lifetime of the preview session. Any ambiguity in the intent must be resolved before proceeding.

TESTNET_ONLY

This skill operates exclusively on Monad Testnet (chain ID 10143). If the RPC responds with a different chain ID the request must be immediately blocked. Never use a mainnet RPC endpoint. The RPC URL is a server-side constant; it must never be derived from user input (SSRF protection).

DECIMAL_STRINGS

All token amounts MUST be represented as decimal strings (e.g. "1.5", "0.001"). Never use floating-point numbers for amounts — they are subject to precision loss. Parse amounts to bigint via parseEther only at the point of RPC call construction.

NO_PRIVATE_KEYS

This skill must never request, store, process, or transmit private keys, seed phrases, or any credential that controls a wallet. All operations are read-only from the perspective of key material.

NO_SIGNING

This skill constructs unsigned transaction objects for preview purposes only. It must never call any signing function (eth_sign, personal_sign, eth_signTypedData, wallet_signTransaction, etc.) or any library method that requires a private key to sign.

NO_BROADCAST

This skill must never submit transactions to the network. Methods such as eth_sendRawTransaction, eth_sendTransaction, or any library wrapper that broadcasts are strictly forbidden. The unsigned transaction object is returned to the caller for display only.

SIMULATION_REQUIRED

A preview_simulate call with live on-chain evidence (balance confirmation, gas estimate, block timestamp) must complete successfully before the result is presented to the user. The simulation result becomes part of the returned artifact. A preview without simulation evidence must be treated as BLOCKED.

STOP_ON_WARNING

If any preflight check yields a warning — chain ID mismatch, RPC timeout, invalid address, zero address, non-positive amount, insufficient balance, gas estimation failure, or missing simulation — the decision must be set to BLOCKED and processing must stop. There is no concept of a "soft warning" that allows the flow to continue.

PRESENT_BEFORE_SIGNING

The full preview artifact (network evidence, unsigned transaction, gas estimate, decision, all safety flags) must be presented to the user for review before any wallet interaction is initiated. The agent must not skip the review step or proceed autonomously to wallet submission.

Data Flow

preview_discover → preview_load → preview_action → preview_simulate
  1. preview_discover — returns the available action descriptor.
  2. preview_load — returns the input contract (required fields, types) and the active safety rule IDs.
  3. preview_action — validates addresses and amount client-side, constructs an unsigned transaction object. No RPC call at this step.
  4. preview_simulate — requires the output of preview_action. Queries live Monad Testnet RPC for: chain ID (verified against 10143), latest block number and timestamp, sender balance, gas estimate, gas price, and optionally recipient code (to flag contract recipients). Returns network evidence and a final decision.

Decision Values

  • READY_FOR_WALLET_REVIEW — all preflight checks passed; safe to present to the user for wallet submission.
  • BLOCKED — one or more preflight checks failed; do not proceed to wallet.

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.