The Agent Skills Directory

[SAFE]: No security issues detected. The skill provides search and fetch capabilities using a documented API service.
[DATA_EXPOSURE]: The skill documentation describes how to read credentials from the platform's standard configuration file (~/.gooseworks/credentials.json). This is a documented setup procedure for the ecosystem and does not represent an exfiltration attempt.
[EXTERNAL_DOWNLOADS]: The skill references standard package runners (npx) for installation and authentication using toolsets associated with the platform (gooseworks, goose-skills).
[INDIRECT_PROMPT_INJECTION]: The skill provides the ability to ingest data from arbitrary external URLs via the /fetch endpoint.
Ingestion points: URL parameter for the /fetch endpoint in SKILL.md.
Boundary markers: Not present in the current prompt instructions.
Capability inventory: Uses curl to send data to the gooseworks.ai proxy API.
Sanitization: No explicit sanitization or filtering is defined in the instructions for content fetched from remote URLs.

web-search-linkup