Security audit
Security checks for vulnerabilities and agentic risk
This skill is transparently about paying people through Talentir, but it gives an agent high-impact payment and webhook-changing abilities without enough built-in user-confirmation guidance.
Install only if you intend to let the agent work with Talentir payments, and use a least-privilege API key. Do not grant payout approval or webhook administration unless needed, require human confirmation for every recipient, amount, currency, and webhook target URL, and treat returned webhook signing secrets as sensitive credentials.
The skill enables real financial payouts and webhook creation/deletion without strong warnings about irreversible or externally visible consequences. In an agent setting, that creates a meaningful risk of accidental fund transfers, unauthorized recipient payments, or webhook changes that alter downstream system behavior.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
All requests need:
curl -s "https://www.talentir.com/api/v1/..." \
-H "Authorization: Bearer $TALENTIR_API_KEY" \
-H "Content-Type: application/json"
This endpoint transmits recipient identifiers and triggers payout creation on an external service, making it a state-changing financial operation rather than benign data transfer. If invoked by an agent without strict authorization and confirmation, it can directly cause unauthorized payments to attacker-chosen social handles.
Supported platforms: tiktok, instagram, youtube-channel.
curl -s -X POST "https://www.talentir.com/api/v1/payout" \
-H "Authorization: Bearer $TALENTIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
This variant also creates payouts and includes tags/custom IDs that can make automated or bulk financial actions easier to operationalize. The danger is the same core issue: authenticated external requests can move money or initiate obligations if an agent is tricked or misconfigured.
curl -s -X POST "https://www.talentir.com/api/v1/payout" \
-H "Authorization: Bearer $TALENTIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
The skill is presented primarily as a payment utility, but it also documents webhook creation and deletion, which expands its operational scope into outbound event delivery and infrastructure management. That mismatch can mislead users and agents about available capabilities, increasing the chance of unintended configuration changes or data exfiltration via attacker-controlled webhook endpoints.
Creating webhooks sends account events to a specified external URL and returns a signing secret, so misuse can establish persistent exfiltration or unauthorized integrations. In agent contexts, an attacker-controlled targetUrl could siphon payout metadata or alter system trust boundaries without obvious immediate symptoms.
curl -s -X POST "https://www.talentir.com/api/v1/webhook" \
-H "Authorization: Bearer $TALENTIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
The documented preApproved/auto-approve capability reduces human oversight for a financial action, which is risky when used by autonomous agents. If the permission is enabled, an agent could create and approve payouts in one flow, increasing the chance of fraud, prompt-injection-induced spending, or accidental disbursements.
| `tags` | No | Array of strings for categorization |
| `customId` | No | Your own identifier for the payout |
| `notifications` | No | `allowed` (default) or `not-allowed` |
| `preApproved` | No | `true` to auto-approve (requires `payout.api_approve` permission) |
## Payout Statuses
The stated purpose is to compensate people using Talentir. The included team endpoint retrieves account/team information, which is an account-inspection capability not described in the manifest's payment-focused scope.
No suspicious patterns detected.