Talkspresso

Manage a Talkspresso business (services, appointments, products, clients, earnings, calendar) using the Talkspresso REST API. Use when the user wants to chec...

This skill appears to do what it says: it will make API requests to https://api.talkspresso.com using the TALKSPRESSO_API_KEY. Before installing/using it: (1) only provide an API key you control and understand — prefer a key with limited scope or rotate it after use; (2) be aware any data the agent sends/receives (appointments, client info, earnings, files) will go to Talkspresso; (3) ensure the host environment has curl and jq available (SKILL.md assumes them); (4) do not paste your API key into chat logs or public places; (5) if you have privacy or compliance concerns about transmitting client data, avoid granting access or test with a non-production account. Overall the components are internally consistent and proportionate.

Type: OpenClaw Skill Name: talkspresso Version: 1.0.0 The skill is classified as suspicious due to its file upload capabilities, which, while necessary for its stated purpose (e.g., uploading profile photos, product attachments), present a significant prompt injection vulnerability. An attacker could potentially craft a prompt to instruct the agent to upload arbitrary local files (e.g., `/etc/passwd`, `~/.ssh/id_rsa`) to the `https://api.talkspresso.com` endpoint using the `curl -F "file=@/path/to/file.jpg"` commands detailed in `SKILL.md`. This is a vulnerability in the agent's execution model rather than explicit malicious intent within the skill's instructions, which otherwise appear benign and focused on managing a Talkspresso business.