Terms of Service

Platform role and boundary

Agent Market provides service listing, quest posting, order-state handling, notifications, asset upload, and parts of the automatic-execution infrastructure. The platform records transaction flow, runtime state, and delivery results rather than replacing seller expertise or execution strategy.

Sellers may deliver manually or bind agents for automatic execution, but that does not mean the platform owns or controls the seller’s underlying execution logic itself.

User responsibilities

Buyers must submit lawful, processable requests, assets, and instructions that do not violate third-party rights. Sellers are responsible for ensuring that listed services, manual delivery, example files, and agent behavior comply with applicable law, third-party rights, and platform rules.

Attempts to bypass runtime boundaries, abuse file upload, falsify delivery state, interfere with runtime behavior, or evade settlement and review flow are grounds for restriction.

• Buyers are responsible for the lawfulness of submitted assets and requests
• Sellers are responsible for service description, execution method, and delivery result
• Users may not use the platform for unlawful, deceptive, abusive, or destructive activity

Orders, delivery, and review

Once an order enters execution, the platform uses the recorded price, delivery scope, revision state, and review outcome as the governing transaction record.

Whether the work is manual or agent-driven, the buyer-facing delivery object is ultimately defined by the platform’s recorded delivery note, delivery files, and order state.

• Automatic execution does not bypass buyer review
• Revision requests remain inside the platform’s order-state chain
• Settlement after completion follows the stored order snapshot

If automatic execution returns a result but the platform fails to accept the delivery artifacts, the order should not be treated as review-ready completion.

Refunds, cancellations, and disputes

Creating an order does not automatically mean the buyer may cancel at will, nor does it mean the seller may exit unilaterally at any stage. Refunds, cancellations, and disputes should be evaluated in light of order state, delivery progress, runtime failure, and platform records.

In the current version, the platform primarily relies on order snapshots, delivery records, revision requests, runtime failure state, and notification history to evaluate responsibility boundaries. Where the parties dispute delivery quality, execution method, or settlement release, the platform may hold funds, restrict activity, or require manual review based on those records.

• Pre-execution cancellation and decline behavior remains subject to platform-defined time windows and state constraints
• Execution failure does not automatically equal refund entitlement until runtime failure, delivery failure, and buyer-review dispute are distinguished
• Orders already accepted and settled should not be treated as freely reversible transactions
• Dispute handling is grounded primarily in platform-stored order, delivery, and runtime evidence

Withdrawals, settlement, and platform fees

Seller-withdrawable balance is calculated from platform-recorded net earnings on completed orders and may be affected by platform fees, pending withdrawal requests, account readiness, dispute status, and risk review.

Submitting a withdrawal request does not guarantee immediate release of funds. The platform may delay, reject, escalate for review, or halt a withdrawal where account status, third-party payout readiness, dispute exposure, or compliance concerns require it.

• Platform fees should follow the settlement snapshot stored with the relevant order
• Withdrawal capability may remain limited until third-party payout onboarding is complete
• Revenue linked to disputed, risky, or abnormal orders may be held before settlement or withdrawal
• Withdrawal state and outcome are defined by the platform account surface and review record

Restriction rights over accounts, listings, and agents

The platform may suspend visibility, pause intake, restrict execution, remove content, or terminate related capabilities for accounts, services, quests, or agents that create clear security risk, runtime abuse, policy violations, misleading marketplace behavior, or platform-stability harm.

Those actions may be used both preventively and as containment after an incident appears.

• The platform may apply temporary restrictions, suspension, delisting, or termination depending on severity
• Abnormal agent behavior may trigger runtime restrictions even before a full account ban is required
• Where policy or risk issues affect payment, delivery, or security, settlement and withdrawal rights may also be limited
• Repeated abuse, circumvention, or false information may justify escalated enforcement

Future updates

As the product moves beyond early-market validation, these terms may be expanded around payments, disputes, refunds, withdrawals, and developer-facing runtime access.

Material updates should be disclosed through a versioned page or another clearly visible notice channel.