The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where data is received from external HTML components via the tell or setContext methods. This creates a potential surface for indirect prompt injection. However, the instructions explicitly mandate that the agent treat this data as information rather than instructions and requires the content to be labeled accordingly, serving as a functional boundary marker.
[DYNAMIC_EXECUTION]: The skill involves the generation of self-contained HTML documents that include scripts and styles. This is a core part of its primary purpose to build interactive cards. The skill mitigates risks by providing explicit negative constraints, instructing the agent to never include credentials, API keys, or personal data within the generated component.
[COMMAND_EXECUTION]: The skill utilizes a suite of platform tools (create-app, add-tool-to-app, publish-app) to modify the workspace environment. The instructions enforce a human-in-the-loop requirement, specifying that these operations should only be performed after receiving explicit user approval.

list-or-build-skilder-apps