This release adds 2 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Summary
AI summaryThinking Tab Color Fix improves readability, Config Improvements auto-migrate old configs, and Previous Fixes Included enhance stability.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Bugfix | Medium |
Thinking tab text now uses muted color for distinction Thinking tab text now uses muted color for distinction Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Config loading gracefully handles missing provider/model/base_url fields with defaults Config loading gracefully handles missing provider/model/base_url fields with defaults Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Model selection persists across restarts Model selection persists across restarts Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Step count no longer shows previous task's steps Step count no longer shows previous task's steps Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Expand/collapse label text corrected Expand/collapse label text corrected Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Agent always shows output, eliminating empty responses Agent always shows output, eliminating empty responses Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Refactor | Low |
Branding updated from previous names to Terminator/OpenSkynet Branding updated from previous names to Terminator/OpenSkynet Source: llm_adapter@2026-06-03 Confidence: high |
— |
Full changelog
What's Fixed
Thinking Tab Color Fix
- Problem: Thinking text was same color as response, making it hard to distinguish
- Solution:
- Thinking/Planning text now uses muted color (text_muted)
- Responding text uses regular text color
- Executing text uses regular text color
- Result: Easy to distinguish between thinking phase and actual response
Config Improvements
- Problem: Old config files missing provider/model/base_url fields caused issues
- Solution: Config loading now handles missing fields gracefully with defaults
- Result: Old configs auto-migrate to new structure
Previous Fixes Included
- Model selection now persists across restarts
- Step count no longer shows previous task's steps
- Expand/collapse label text fixed
- Agent always shows output (no empty responses)
- Branding updated to Terminator/OpenSkynet
Installation
curl -fsSL https://raw.githubusercontent.com/sediman-agent/OpenSkynet/main/scripts/install.sh | bash
For aarch64-macos (Apple Silicon):
Pre-built binaries not available. Build from source:
git clone https://github.com/sediman-agent/OpenSkynet.git
cd OpenSkynet
cargo build --release --package sediman-tui
cp target/release/sediman-tui ~/.local/bin/terminator
chmod +x ~/.local/bin/terminator
Manual Config Update (if needed)
If your config is missing provider/model, add to ~/.terminator/tui.toml:
provider = "minimax"
model = "m3"
base_url = ""
Full Changelog: https://github.com/sediman-agent/OpenSkynet/compare/v0.3.6...v0.3.7
Full Changelog: https://github.com/sediman-agent/OpenSkynet/compare/v0.3.6...v0.3.7
Full Changelog: https://github.com/sediman-agent/OpenSkynet/compare/v0.3.6...v0.3.7
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Sediman Browse
All releases →Related context
Related tools
Earlier breaking changes
- v0.2.7 Renamed data directory from ~/.sediman to ~/.terminator across Python and Rust codebases
Beta — feedback welcome: [email protected]