demo-VM e2e validation of Phase 11 (service_agent_v3 + Python flow cutover) #23

New issue

Open

opened 2026-05-06 17:07:25 +00:00 by mik-tf · 0 comments

mik-tf commented

2026-05-06 17:07:25 +00:00

Owner

Overview

#21 (Phase 11-D — DAG deletion) and #22 (Phase 11-E — seed service_agent_v3) are both green and mergeable as of 2026-05-06. The PR-22 test plan still has one unchecked item: manual end-to-end with hero_proc + hero_aibroker + hero_osis_calendar on a real VM. Run that validation, then merge.

Why

Phase 11 is the substrate for the AI agent driving the OS via Python flows. Without VM-level e2e, "Timur Python work" hasn't actually proven the agent loop works.

Acceptance

Deploy #21 + #22 on heroci (or herodemo)
Run cargo test -p hero_logic --test e2e_create_event against live aibroker + osis_calendar — assert pass
Manually invoke service_agent_v3 with a real prompt, check Play.spans tree populated
Capture timing + Span output in a comment here
Squash-merge after green

hero_logic#14 — sandbox roadmap parent
hero_logic#21
hero_logic#22

Source: meeting notes 2026-05-06.

## Overview [#21](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/21) (Phase 11-D — DAG deletion) and [#22](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/22) (Phase 11-E — seed service_agent_v3) are both green and mergeable as of 2026-05-06. The PR-22 test plan still has one unchecked item: **manual end-to-end with `hero_proc` + `hero_aibroker` + `hero_osis_calendar` on a real VM**. Run that validation, then merge. ## Why Phase 11 is the substrate for the AI agent driving the OS via Python flows. Without VM-level e2e, "Timur Python work" hasn't actually proven the agent loop works. ## Acceptance - [ ] Deploy [#21](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/21) + [#22](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/22) on heroci (or herodemo) - [ ] Run `cargo test -p hero_logic --test e2e_create_event` against live aibroker + osis_calendar — assert pass - [ ] Manually invoke service_agent_v3 with a real prompt, check Play.spans tree populated - [ ] Capture timing + Span output in a comment here - [ ] Squash-merge after green ## Related - [hero_logic#14](https://forge.ourworld.tf/lhumina_code/hero_logic/issues/14) — sandbox roadmap parent - [hero_logic#21](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/21) - [hero_logic#22](https://forge.ourworld.tf/lhumina_code/hero_logic/pulls/22) Source: meeting notes 2026-05-06.