2.21.0
2.21.0
What's Changed
- Site Improvements for Accessibility and Load Speed by @willbonde in #1004
- Improve eval developer ergonomics by @ai-christianson in #1007
- Add remote MCP agent debug trace tool by @ai-christianson in #1008
- Improve effort calibration and overwork behavior by @ai-christianson in #1009
- Prompt Optimizations from Evals by @matt-greathouse in #1005
- Improve discord integration by @matt-greathouse in #1006
- Improve discord integration by @matt-greathouse in #1011
- update_plan improvements by @matt-greathouse in #1013
- SMS contact approval with documentation by @willbonde in #1015
- Fix remaining eval cases and reduce prompt size by @matt-greathouse in #1012
- Fix human input dismiss resuming agents by @ai-christianson in #1014
- TOS Update and new newsletter by @willbonde in #1018
- Guard custom tool SQLite sync against malformed DBs by @matt-greathouse in #1017
- Prompt updates based on judge by @matt-greathouse in #1016
- Add allowed params for openai-compatible endpoints by @matt-greathouse in #1019
- Custom tool improvements and evals by @matt-greathouse in #1020
- Mobile Performance Improvements by @willbonde in #1010
- Agent prompt and eval improvements by @matt-greathouse in #1022
- Updating library for SEO and Performance by @willbonde in #1023
- Add view all btn for contact requests and fix context switching in live chat by @matt-greathouse in #1026
- Charter fixes and evals by @matt-greathouse in #1027
- SEO - Library Urls by @willbonde in #1025
Full Changelog: 2.20.1...2.21.0