v0.36.4

cortezaproject/cortezav0.36.4Jun 17, 2026by harijoe

AI Summary

A patch release addressing caching issues where ChatGPT's 30-minute cache limit caused view rendering to break. The fix ensures assets are served based on path resolution regardless of the `?v=` query parameter.

Key Highlights

Fix: Serve view resource no matter the `?v=` cache key.
Resources read now resolves views by query-less path.
Version parameter acts only as a cache-busting hint and no longer gates resolution.

New Features

Improved asset serving stability against cache expiration.
Path-based resolution for view resources.

Full Release Notes

## Patch release

  ### Explanation

ChatGPT used to cache the HTML file referenced in a tool resource indefinitely. This recently changed: the cache now lasts about 30min and is then refetched. Before this patch, Skybridge only served the HTML when the exact version hash was provided as a `?v=` query param, so resubmitting an app would break view rendering in all earlier conversations once their cache expired and they refetched the old hash.

  ### Fix
  
`resources/read` now resolves a view by its query-less path, so the underlying asset is served no matter what `?v=` value the consumer sends (stale cache key, mismatched version, or no param at all). The version param is only a cache-busting hint and no longer gates resolution. The consumer-facing URI is preserved on the response.
  
  ## Changes
- fix: serve view resource no matter the ?v= cache key @paulleseute (https://github.com/alpic-ai/skybridge/pull/872)