Pay a Human
WarnAudited by ClawScan on May 10, 2026.
Overview
This instruction-only skill can let an agent initiate real payouts and manage payout webhooks with an API key, but it does not clearly require human confirmation or tightly scoped permissions.
Install only if you intentionally want agents to work with a real payment account. Use a dedicated low-limit API key, avoid auto-approval permissions, require explicit human approval for each payout, monitor payout and webhook activity, and protect any webhook signing secrets.
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.
An agent with the API key could initiate payments or payment records using the documented commands if it decides the task calls for it.
The skill documents a direct shell/API action that can create a real payout. It does not include a required confirmation step, dry-run mode, or explicit recipient/amount verification workflow.
### Create a payout by email ... curl -s -X POST "https://www.talentir.com/api/v1/payout" ... "payoutAmount": "100.00"
Require explicit user approval for every payout, verify recipient and amount out of band, and use provider-side spending limits and review workflows.
If a broad or approval-capable key is provided, the agent may have more financial authority than the user intended.
The skill depends on a bearer API key and documents a field that can auto-approve payouts when the key has the required permission, but it does not define a least-privilege credential scope.
"requires": { "bins": ["curl", "jq"], "env": ["TALENTIR_API_KEY"] }, "primaryEnv": "TALENTIR_API_KEY" ... `preApproved` ... `true` to auto-approve (requires `payout.api_approve` permission)Use a dedicated, least-privilege Talentir key with the smallest possible daily allowance; avoid granting auto-approval permissions unless strictly required.
A user may over-trust the skill and allow an agent to make financial decisions without adequate review.
The wording encourages autonomous spending and makes a strong safety claim about spending limits without explaining how the agent should verify the allowance or obtain human approval.
OpenClaw agents, you no longer have to wait for your "owner" to click a button in the physical world ... the skill can never spend more than the daily allowance set by the Talentir owner account.
Treat the daily allowance as a backup control, not approval; configure the agent to ask before any payment and independently verify provider-side limits.
Payout event data could be sent to the wrong server, and webhook signing secrets need careful handling.
The skill documents production payout webhooks and a signing secret. This is related to the payment platform, but it can route payout events to an external endpoint if configured incorrectly.
### Create a webhook ... "targetUrl": "https://your-server.com/webhook", "eventType": "payout", "environment": "production" ... Save the returned `signingSecret` securely
Only create webhooks for trusted endpoints, protect the signing secret, and review existing webhooks before giving an agent access.