From c31385d7757881ace5be4529ee76dc484911f4d1 Mon Sep 17 00:00:00 2001 From: Ayla Croft Date: Sun, 20 Sep 2026 07:19:56 -0400 Subject: [PATCH] plan: ADR-0009 records the owner's decision on Jido v2 Change 5 of the 2026-09-20 plan. Appended, never rewritten: the runtime is Jido v2 at ~> 2.3 (2.3.3 stable; the v3 line is 3.0.0-beta.1 as of 2026-09-14). The Slice 012 checkpoint still measures what the ADR asked it to and can no longer return "reject". The shared-library note goes to the standards register, not to a slice. plan_check: PASS. Signed-off-by: Ayla Croft --- .../0009-jido-for-the-action-effect-layer.md | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/docs/adr/0009-jido-for-the-action-effect-layer.md b/docs/adr/0009-jido-for-the-action-effect-layer.md index 53e3eef..6e94ce7 100644 --- a/docs/adr/0009-jido-for-the-action-effect-layer.md +++ b/docs/adr/0009-jido-for-the-action-effect-layer.md @@ -1,5 +1,5 @@ # ADR-0009 — Jido 2.0 for actions, directives and the effect boundary; decided by the Slice 012 design checkpoint -Status: proposed · Date: 2026-09-05 +Status: accepted · Date: 2026-09-05 · Owner decision on the version line recorded 2026-09-20 ## Context ADR/tech-stack v1 said "Jido not chosen". Reconsidered: Jido has a worked vocabulary for actions, directives and @@ -19,3 +19,18 @@ finalised. - Slice 020's tool behaviour becomes a thin wrapper over `Jido.Action` if adopted; the registry rules (compile-time catalog, runtime tools cannot enter it) are unchanged. - VERSIONS.md gains a `jido ~> 2.0` row pending the checkpoint. + +## Decision, appended 2026-09-20 + +Owner decision, before 2026-09-19, recorded here on 2026-09-20: **the runtime is Jido v2**, pinned `~> 2.3` +(2.3.3 on hex.pm as of this record; the v3 line is at 3.0.0-beta.1, 2026-09-14, and defers live code migration +and distributed control-plane claims by its own release notes). The provisional decision above is confirmed on +the version question and on the shape: actions, directives and the effect boundary use Jido's vocabulary; the +Session stays an OTP `gen_statem`; the compile-time effect catalog and the rule that runtime tools never enter it +are unchanged whichever way `Jido.Agent` lands. The Slice 012 design checkpoint still measures (a), (b) and (c) +as written and reports; what it can no longer return is "reject", because the runtime question is decided. It +may still return "actions and directives only". + +One consequence for the standards register rather than for a slice: another system in the same platform family +runs on the same Jido line. That is a shared library, not a shared runtime, and it is not on the path a finding +takes; the register carries the row and the argument, and no slice here does.