The Agent Skills Directory

[SAFE]: The skill provides documentation for interacting with specific platform APIs (canvas-source-retrieve, canvas-publish-create, canvas-edit-create, canvas-draft-create, etc.) to manage application source code.
[SAFE]: It emphasizes security-centric workflows, such as identifying and declaring network origins, insights, and event capture capabilities before publishing. It explicitly notes that private/local network origins are rejected by the host.
[SAFE]: It describes and provides guidance on addressing security diagnostics from the platform's validation engine, which blocks patterns such as forbidden dynamic imports, unauthorized bare imports, and inline scripts.
[SAFE]: The publishing and promotion workflows include safety mechanisms like version guards (expected_current_version_id) to prevent race conditions and overwriting concurrent work.

validating-and-publishing-canvases