Tracking the post-v0.3.1 6-item plan documented in docs/ROADMAP.md. Sequence:
4/6 done in 2 days (target was 3 months). Remaining items are open-ended (recipes can pick 3-5 from a candidate list; phase 2+3 is real systems engineering).
Rationale
Phase 1 source-pause work shipped at 143× ceiling / 6-15× typical / 14× multi-BRANCH aggregate (RESULTS-v0.3.md). Pushing the number further past 200 ms control-plane floor is diminishing returns until real user workloads point at the next bottleneck.
Items 1-3 maximize distribution (MCP for any agent client, GitHub Action for CI flow, TS SDK for JS-side agent code). Item 4 closed the Daytona stickiness gap. Items 5-6 (recipes, technical depth) come once 1-4 produce user signal.
Track classifications
Tracking the post-v0.3.1 6-item plan documented in docs/ROADMAP.md. Sequence:
@deeplethe/forkd 0.3.1). npm publish pending NPM_TOKEN setup by the maintainer.4/6 done in 2 days (target was 3 months). Remaining items are open-ended (recipes can pick 3-5 from a candidate list; phase 2+3 is real systems engineering).
Rationale
Phase 1 source-pause work shipped at 143× ceiling / 6-15× typical / 14× multi-BRANCH aggregate (RESULTS-v0.3.md). Pushing the number further past 200 ms control-plane floor is diminishing returns until real user workloads point at the next bottleneck.
Items 1-3 maximize distribution (MCP for any agent client, GitHub Action for CI flow, TS SDK for JS-side agent code). Item 4 closed the Daytona stickiness gap. Items 5-6 (recipes, technical depth) come once 1-4 produce user signal.
Track classifications