Skill

Manages travel for your user via UBTRIPPIN — trips, items, loyalty programs, family, city guides, events, concerts, notifications, and more. Use when the use...

Before installing, verify the ubtrippin service and publisher (check ubtrippin.xyz and the referenced GitHub repo). Be aware the skill requires you to provide a UBTRIPPIN API key (ubt_k1_...) and a registered sender email — these are not declared in the registry metadata, so treat that as a transparency gap. Consider these safety steps: (1) only supply an API key created specifically for this integration and keep it scoped/rotatable if the service supports it; (2) avoid storing the key in plaintext in long-lived agent configs; use ephemeral tokens or a secrets manager if possible; (3) understand that forwarding booking emails (with attachments) to [email protected] will transmit sensitive travel and identity data — confirm you trust the destination and its privacy policy; (4) ask the publisher to correct the registry metadata to declare required env vars and primary credential; (5) test with non-sensitive/demo data first, and be ready to revoke the API key if you see unexpected behavior.

Type: OpenClaw Skill Name: ubtrippin Version: 2.3.0 The skill bundle is classified as suspicious due to the presence of high-risk capabilities that could be exploited via prompt injection against the AI agent. Specifically, the `SKILL.md` file documents a `webhooks` API (`POST /api/v1/webhooks`) which allows the agent to configure an arbitrary URL to receive real-time event data (e.g., `trip.created`, `item.added`). This presents a significant data exfiltration risk, as a compromised agent could be instructed to send sensitive user travel data to an attacker-controlled server. Additionally, the `GET /api/v1/me/loyalty/export` endpoint allows for bulk export of sensitive loyalty program data, which could also be exfiltrated if the agent is maliciously prompted. These are legitimate API features, but their exposure to an AI agent creates a critical vulnerability.