[4.15.0]
Pulled in all changes from Roo Code up until v3.11.17
[3.11.17] - 2025-04-14
Improvements to OpenAI cache reporting and cost estimates (thanks @monotykamary and Cline!)
Visual improvements to the auto-approve toggles (thanks @sachasayan!)
Bugfix to diff apply logic (thanks @avtc for the test case!) and telemetry to track errors going forward
Fix race condition in capturing short-running terminal commands (thanks @KJ7LNW!)
Fix eslint error (thanks @nobu007!)
[3.11.16] - 2025-04-14
Include model ID in environment details and when exporting tasks (thanks @feifei325!)
[3.11.15] - 2025-04-13
Add ability to filter task history by workspace (thanks @samhvw8!)
Fix Node.js version in the .tool-versions file (thanks @bogdan0083!)
Fix duplicate suggested mentions for open tabs (thanks @samhvw8!)
Fix Bedrock ARN validation and token expiry issue when using profiles (thanks @vagadiya!)
Add Anthropic option to pass API token as Authorization header instead of X-Api-Key (thanks @mecab!)
Better documentation for adding new settings (thanks @KJ7LNW!)
Localize package.json (thanks @samhvw8!)
Add option to hide the welcome message and fix the background color for the new profile dialog (thanks @zhangtony239!)
Restore the focus ring for the VSCodeButton component (thanks @pokutuna!)
[3.11.14] - 2025-04-11
Support symbolic links in rules folders to directories and other symbolic links (thanks @taisukeoe!)
Stronger enforcement of the setting to always read full files instead of doing partial reads
[3.11.13] - 2025-04-11
Loads of terminal improvements: command delay, PowerShell counter, and ZSH EOL mark (thanks @KJ7LNW!)
Add file context tracking system (thanks @samhvw8 and @canvrno!)
Improved display of diff errors + easy copying for investigation
Fixes to .vscodeignore (thanks @franekp!)
Fix a zh-CN translation for model capabilities (thanks @zhangtony239!)
Rename AWS Bedrock to Amazon Bedrock (thanks @ronyblum!)
Update extension title and description (thanks @StevenTCramer!)
[3.11.12] - 2025-04-09
Make Grok3 streaming work with OpenAI Compatible (thanks @amittell!)
Tweak diff editing logic to make it more tolerant of model errors