Freelance Automation Gig
MaliciousAudited by ClawScan on May 1, 2026.
Overview
This skill explicitly tells the agent to pose as a human freelancer while operating on freelance marketplaces, messaging clients, and handling payments or disputes.
Avoid installing this skill as written. It is designed to let an AI agent act in freelance marketplaces while pretending to be human, and it gives broad instructions for client messaging, automation work, invoices, payment checks, and disputes without clear approval safeguards.
Findings (4)
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.
Using this skill could mislead clients and platforms about who is doing the work, creating trust, policy, contractual, and reputational risk.
This explicitly instructs the agent to conceal that it is AI and pretend to be a human freelancer, which is direct deception toward clients or marketplace participants.
絕對唔好透露自己係 AI（扮 human freelancer）
Do not install or use the skill as written; require transparent disclosure, platform-compliant representation, and human review before any client-facing action.
The agent could act through a user's freelancer account in ways that affect money, contracts, client relationships, and marketplace standing.
The skill directs the agent toward marketplace account, invoice, payment-method, and dispute actions, but the registry declares no primary credential or scoped account-access contract.
Invoice Generation - 自動開發票 ... 檢查 Upwork/Fiverr payment method，必要时 disputed
Require explicit account authorization, narrow scopes, per-action user approval, and clear rules for payments, invoices, disputes, and client communications.
The agent may perform broad external automation that can violate site rules, send unwanted messages, or create systems that affect third parties.
The skill authorizes broad automation work, including scraping, email automation, and bot creation, without defining target limits, legality checks, rate limits, approval gates, or safe-use boundaries.
Task Execution - 執行各種自動化任務 ... Web Scraping ... Email Automation ... Bot Creation
Limit execution to user-approved targets and tasks, require compliance checks, and add explicit confirmation before scraping, emailing, bot deployment, or other external actions.
The agent could continue contacting clients later, potentially without the user re-confirming the message, timing, or context.
The skill instructs delayed follow-up messaging after 48 hours, implying ongoing autonomous activity beyond a single immediate user request.
"Client not responding" → 發 follow-up message，48小時後再發一次
Require explicit scheduling consent, message preview, cancellation controls, and user confirmation before each client-facing follow-up.