Funpay Assistant

WarnAudited by ClawScan on May 10, 2026.

Overview

The skill matches its FunPay chat-automation purpose, but it ships a hard-coded FunPay account key, previous chat/order data, and code that can automatically message customers about granting account access.

Do not run this skill as-is. Treat the embedded FunPay key as compromised, remove state.json before distribution, require user-provided credentials, and add explicit review before any outgoing customer message or account-access request.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Anyone running or receiving the skill may expose or act through the FunPay account tied to that key, including reading chats and sending messages as that account.

Why it was flagged

The script embeds a FunPay authentication key and uses it to log in, read chats, and send messages, while the registry metadata declares no primary credential.

Skill content
GOLDEN_KEY = "r7sb47...j02bz" ... acc = Account(GOLDEN_KEY).get()
Recommendation

Revoke and rotate the embedded key, remove it from the package, and require each user to provide their own scoped credential through declared configuration.

What this means

A single run can send automated business messages to multiple customers, including inappropriate replies caused by broad keyword matches.

Why it was flagged

The code loops over all FunPay chats returned by the account and sends replies based on keyword matches without asking the user to approve each outgoing marketplace message.

Skill content
chats = acc.request_chats() ... for chat in chats: ... safe_send_message(acc, chat.id, REGION_ERROR_REPLY)
Recommendation

Add a dry-run mode, per-message approval, chat/order allowlists, and clear logs before sending any FunPay message.

What this means

Customers could be automatically encouraged to grant account access, which is a high-trust and high-impact action even if it is part of the seller's workflow.

Why it was flagged

The automated reply proposes that the seller log into a buyer's account, and the script later flags messages like consent or 'go ahead' as urgent.

Skill content
могу предложить такой вариант: Я захожу к вам на аккаунт и активирую ваш ключ
Recommendation

Avoid automating requests for third-party account access; require human review and use safer activation methods that do not require customers to share or delegate account access.

What this means

Private marketplace conversation and order history can be exposed to anyone who obtains the skill, and the carried-over state can affect future bot behavior.

Why it was flagged

The package includes persistent chat IDs, message contents, buyer/order references, and auto-reply state from prior FunPay activity.

Skill content
"last_message_texts": { "239122985": "К сожалению...", ... "Покупатель ... подтвердил успешное выполнение заказа #..." }
Recommendation

Do not publish live state files; remove existing chat history, store only minimal local state, and define retention, redaction, and reset controls.

What this means

Users cannot verify from these artifacts exactly which dependency implementation will handle FunPay login and messaging.

Why it was flagged

The skill relies on an external FunPayAPI library, but the provided artifacts include no install spec, lockfile, source, or dependency version for that account-access code.

Skill content
from FunPayAPI.account import Account
Recommendation

Declare and pin dependencies, provide a trusted source or lockfile, and document the exact installation path before release.