Agent Ads Earn
WarnAudited by ClawScan on May 10, 2026.
Overview
This skill openly monetizes monitoring XMTP group chats and sending human identity/chat-context referrals to a remote ad marketplace, but the artifacts do not clearly bound consent, privacy, retention, or authorization.
Review this skill carefully before installing. It is not just an earnings helper; it monitors XMTP group conversations for commercial intent and can send information about specific people and messages to a remote referral service. Use it only in groups where participants understand and accept that monitoring/referral may occur, and confirm the provider's authentication, retention, deletion, and anti-spam policies.
Findings (6)
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.
People in monitored chats could have identifying information and message context sent to an advertising/referral platform before they have agreed to join the advertiser's group.
The referral API can transmit a person's messaging identifier, wallet address, source group, interests, and the triggering chat message to a remote service.
"humanInboxId": { "type": "string", "description": "XMTP inbox ID of the matched human" }, "humanWallet": { "type": "string", "description": "Wallet address of the matched human" }, ... "triggerMessage": { "type": "string", "description": "The message that triggered the match (max 500 chars)" }Install only if group members are informed and have opted in, and ask the provider how referral data is stored, retained, shared, and deleted.
A false match or overly aggressive configuration could cause unsolicited referral DMs or ad-related outreach to people in your chats.
The core workflow tells the agent to continuously inspect group chat messages for monetizable intent.
Monitor messages in your group chats for intent signals that match subscriber interests.
Require explicit user approval before each referral, set a high confidence threshold, and limit monitoring to groups where members know this is happening.
The agent may prioritize paid referral opportunities over the best interests or expectations of people in the chat.
The skill encourages optimizing intent detection around payout rates, creating a financial incentive to route humans to advertisers.
Higher `avgCphRate` = more earning potential.
Use clear disclosure in monitored groups and do not allow the agent to refer people based only on payout value.
If publisher IDs are not protected by additional server-side authorization, another party could potentially query earnings or submit referrals under someone else's publisher identity.
The artifact says later calls use only a publisher ID, while referral submission and dashboard access are tied to that identifier.
"authentication": { "type": "erc-8004", "description": "Publisher registration requires ERC-8004 identity on Base. Subsequent calls use publisher ID." }Confirm that the service requires signed wallet/XMTP authentication for dashboard and referral actions, not just possession of a publisher ID.
Monitoring can continue for registered groups until they are removed or the publisher setup is changed.
The service maintains an active list of groups to monitor, which is persistent behavior, although it is aligned with the skill's disclosed purpose.
"groupIds": { "type": "array", "items": { "type": "string" }, "description": "XMTP group IDs to monitor for intent" }Review registered groups periodically and remove any group that should no longer be monitored.
You must trust the remote Basemate endpoint's behavior without seeing its implementation in these artifacts.
There is no local package to inspect and the source is not identified, while the skill relies on a remote service for sensitive referral functionality.
Source: unknown; Homepage: none; No install spec — this is an instruction-only skill.
Verify the provider and endpoint out of band before sending chat or wallet-related data.