The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill retrieves and displays untrusted data that could contain instructions designed to manipulate the AI's behavior.\n
Ingestion points: The handleAgent function in index.ts ingests agent.description and agent.personality from an external API.\n
Boundary markers: Absent. Untrusted strings are appended directly to the output without delimiters or warnings to ignore embedded instructions.\n
Capability inventory: The skill's functionality is limited to data display and formatting, lacking dangerous capabilities like system command execution or file system writes.\n
Sanitization: Absent. While descriptions are truncated to 300 characters, there is no filtering or escaping of the content.

virtuals