The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from ezviz.com, which is an external and untrusted source that could contain hidden instructions targeting the agent.
Ingestion points: Web content is ingested into the agent's context through tools defined in SKILL.md such as ezviz_com__get_search, ezviz_com__read_page, unbrowse.scrape, and unbrowse.run.
Boundary markers: The instructions do not define any specific delimiters or guidance for the agent to distinguish between its system instructions and the data retrieved from the website.
Capability inventory: The skill possesses network capabilities (proxied via the Unbrowse MCP server) and can potentially perform follow-up actions based on the content of the pages it reads.
Sanitization: There is no evidence of content filtering or sanitization being applied to the data retrieved from the external site before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill depends on a remote MCP server and interacts with external vendor infrastructure.
Remote host: The skill configuration and tool logic are hosted on https://unbrowse.ai/mcp/ezviz.com.
Interaction: The skill facilitates communication between the agent and the Unbrowse service for replaying ezviz.com's first-party APIs.

ezviz