The Agent Skills Directory

[SAFE]: The skill consists entirely of architectural documentation and design guidelines for hosted agent infrastructure. It does not contain executable scripts, commands, or automation logic.
[INDIRECT_PROMPT_INJECTION]: The skill describes an architecture that processes input from various external sources (Slack, web interfaces, Chrome extensions). It correctly identifies the necessity for sandbox isolation and tool-call interception to mitigate risks associated with untrusted data processing.
[EXTERNAL_DOWNLOADS]: The skill includes reference links to official documentation and repositories from well-known technology providers and cloud services, such as Cloudflare, Modal, and SST. These references are provided as legitimate educational resources.

hosted-agents