The Agent Skills Directory

[Data Exposure & Exfiltration] (LOW): The script scripts/search.js performs network operations using fetch() to access the non-whitelisted domain www.weavefox.cn and follows arbitrary URLs returned by that API. While no sensitive local data is targeted, these outbound connections are noted.
[Indirect Prompt Injection] (LOW): The skill ingests untrusted content from an external service, which could potentially contain malicious instructions embedded in SVG metadata or comments.
Ingestion points: scripts/search.js fetches data from the Weavefox API and subsequent icon URLs.
Boundary markers: Absent; the raw SVG content is returned to the agent without delimiters or warnings.
Capability inventory: The skill is restricted to read operations and console output; it lacks the ability to write files or execute system commands.
Sanitization: Absent; the script does not sanitize or filter the content retrieved from remote URLs.

icon-retrieval