Repository navigation
B8 (EPIC): full orchestrator dissolution #115
Description
What to build
Full orchestrator dissolution: move ALL control flow (experiment loop + develop-mode DAG walk) into the harness/score, and retire orchestrator.md on the amicode side. Claude Code keeps its root-orchestrator until the harness reaches parity (dual-runtime is correct until then, W-8). Epic — break into sub-issues when B4 lands. (Spec spec-20260708-112732 §1, §3.1, §8.)
Acceptance criteria
- experiment loop + develop DAG both run from harness/score on the amicode runtime
- orchestrator.md no longer carries control flow (amicode side)
- convergence trigger documented; CC root-orchestrator retirement is a follow-up
Blocked by
Provenance
Design spec (Aaron's vault): spec-20260708-112732; plan plan-20260708-214610 (slice B8, EPIC — subdivide when B4 lands).
Activity
aarontrowbridge commentedon Jul 9, 2026
Decomposed (B4 landed, PR #120). Per this epic's own note, subdividing now: #121 (N-iteration loop), #122 (develop DAG walk), #123 (retire orchestrator.md), #124 (real leaf dispatch). All build on the B4 prototype + the merged tool spine, and want Jack's review of the harness approach before implementation. This epic tracks them.
Rchari1 commentedon Aug 20, 2026
Superseded by #368 Autoresearch Studio — the loop is the product (design-of-record `spec-20260813-162300`, decomposed into WS1–WS7 #371–#375).
The `B*` depth-1 redesign (`spec-20260708-112732`) gated by B0 #107 was never ratified; the harness axis has moved to opencode open + plan-first, telaio by subscription (#398) with:
This `B*` slice has had no activity since 2026-07-09 and is now tracked under the WS epics. Closing as superseded — please reopen against #368/WS* if any slice still applies.