The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from udemy.com, creating a surface for potential indirect prompt injection attacks.
Ingestion points: The udemy_com__read_course_by_name tool described in SKILL.md retrieves text and links from external web pages.
Boundary markers: There are no explicit instructions or delimiters provided to help the agent distinguish between course data and potential malicious instructions within that data.
Capability inventory: The skill provides powerful web interaction tools like unbrowse.run and unbrowse.browse.act (defined in SKILL.md), which could be misused if the agent is manipulated by external content.
Sanitization: The instructions do not describe any sanitization or filtering of the content retrieved from Udemy before it is presented to the agent.
[EXTERNAL_DOWNLOADS]: The skill utilizes remote components and installation methods targeting the author's infrastructure.
Evidence: The skill configuration in server.json and instructions in README.md and SKILL.md point to https://unbrowse.ai/mcp/udemy.com. This is the official infrastructure for the skill author, 'unbrowse-ai'.

udemy