github Gentleman-Programming/gentle-ai v1.8.11

latest releases: v1.43.2, v1.43.1, v1.43.0...
3 months ago

Released v1.8.11.

  • Orchestrator delegation rules rewritten around one principle: does this inflate my context without need? The five fragmented sections (Hard Stop, Delegate-First, Anti-Patterns, Task Escalation, Delegation Rules) are replaced by a single table with no room for rationalization.

Read to decide (1-3 files) runs inline. Read to explore (4+ files) delegates. Write atomic (one file, mechanical) runs inline. Write with analysis delegates. Bash state (git, gh) runs inline. Bash execution (test, build) delegates.

The old framing asked the LLM to judge whether something was 'execution' — which LLMs rationalize around. The new framing is a lookup table.

Install: brew update && brew upgrade gentle-ai
Changelog: https://github.com/Gentleman-Programming/gentle-ai/releases/tag/v1.8.11

Don't miss a new gentle-ai release

NewReleases is sending notifications on new releases.