docs: PR #8 follow-ups — Jigo memo, skill stubs, OQ-1, Mnemosyne correction #9

Merged

j7 merged 5 commits from docs/pr8-followups into main

2026-07-10 13:52:30 -04:00

hermes-admin commented

2026-07-10 12:24:14 -04:00

Contributor

What this PR ships

Follow-ups to PR #8 (which was squashed and merged as 2ce34bce). The squash landed Lane A + B1 (11 files, +85/-45), but five follow-up commits landed on the docs/pre-sprint0-hardening branch after the squash was prepared and need a separate PR to reach main:

Commit	What
`62a36b3`	Jigo review memo (imported) + mechanical fixes #2/#3/#7 (Honcho anchor remap, 4 EOF newlines, E2-T1 dedup rationale)
`eba469c`	Jigo review brief (`prompts/jigo-review-pre-sprint0.md`) — needed for Jigo to run reviews on future PRs
`024ed3c`	Epic E5 + E5-T1 seed in `kanban/backlog.md`
`89bbd73`	Skill stubs (E5-T1): `documents/skill-stubs/{README, jigo-review, nixos-host-operations, j7-homelab-environment}.md`
`0736818`	Kohroku Mnemosyne restart correction + Ishii OQ-1 in `docs/architecture/open-questions.md`

Why separate PR

The squash at 2ce34bce did NOT include these commits. Without PR #9, main has the wrong Mnemosyne restart procedure, no open-questions.md, no skill stubs, and no Epic E5 in the backlog.

Verification

Branch: docs/pr8-followups (from main, 5 commits ahead, no conflicts)
Diff vs main: 7 files added, 4 files modified, +294/-7
All commits cherry-picked cleanly from docs/pre-sprint0-hardening; no conflict resolution needed

## What this PR ships Follow-ups to PR #8 (which was squashed and merged as `2ce34bce`). The squash landed Lane A + B1 (11 files, +85/-45), but **five follow-up commits landed on the `docs/pre-sprint0-hardening` branch after the squash was prepared** and need a separate PR to reach `main`: | Commit | What | |---|---| | `62a36b3` | Jigo review memo (imported) + mechanical fixes #2/#3/#7 (Honcho anchor remap, 4 EOF newlines, E2-T1 dedup rationale) | | `eba469c` | Jigo review brief (`prompts/jigo-review-pre-sprint0.md`) — needed for Jigo to run reviews on future PRs | | `024ed3c` | Epic E5 + E5-T1 seed in `kanban/backlog.md` | | `89bbd73` | Skill stubs (E5-T1): `documents/skill-stubs/{README, jigo-review, nixos-host-operations, j7-homelab-environment}.md` | | `0736818` | Kohroku Mnemosyne restart correction + Ishii OQ-1 in `docs/architecture/open-questions.md` | ## Why separate PR The squash at `2ce34bce` did NOT include these commits. Without PR #9, **main has the wrong Mnemosyne restart procedure**, no `open-questions.md`, no skill stubs, and no Epic E5 in the backlog. ## Verification - Branch: `docs/pr8-followups` (from `main`, 5 commits ahead, no conflicts) - Diff vs main: 7 files added, 4 files modified, +294/-7 - All commits cherry-picked cleanly from `docs/pre-sprint0-hardening`; no conflict resolution needed

hermes-admin added 5 commits

2026-07-10 12:24:14 -04:00

docs: post-Jigo minors #2/#3/#7 mechanical fixes + import Jigo review memo 62a36b32f1

- Jigo top-issue #2: docs/persistence-boundaries.md L35 Honcho anchor
  remap (was pointing at non-existent §3 item; Honcho decommission note
  is §2 only).
- Jigo top-issue #3: trailing-newline-at-EOF on kanban/board-schema.md,
  README.md, packets/examples/{backend-kanban-adapter,ui-preview-panel}.md.
- Jigo top-issue #7: E2-T1 dedup row rationale (kept visible for E0
  audit trail after Phase P11 subsumption).
- documents/reviews/: import Jigo review memo for PR #8 per J7 directive
  2026-07-09.

No semantic doc change; A4 Ishii-ack gate unchanged.

prompts: Jigo review brief for PR #8 (pre-Sprint-0 hardening) eba469ca75

kanban: add Epic E5 + E5-T1 (skill stub hardening lane) per Jigo minor #6 024ed3cc69

documents: skill stubs (E5-T1) — pointers for jigo-review, nixos-host-operations, j7-homelab-environment 89bbd73e2c

Per Jigo review of PR #8 (top issue #6): three runtime skills were cited
by canonical docs without in-repo stubs. This commit closes the citation
gap with one-paragraph pointers; full implementations remain canonical
at /home/san/.hermes/skills/{jigo-review,devops/nixos-host-operations,system/j7-homelab-environment}.

Stubs cover purpose / when to load / what it forbids / canonical home.
The README documents the stub vs canonical split so future drift is
caught by Jigo's review lens.

Kanban: t_c9d4a6b2 (E5-T1) status -> in_progress, will mark done after push.

docs: land Kohroku's Mnemosyne restart correction + Ishii's open-questions OQ-1 0736818d79

Two deliverables from Jigo follow-ups landed in workspace by other wolves;
this commit pulls them into PR #8:

- docs/runtime-state.md §3 #1: Mnemosyne restart procedure corrected
  from the (wrong) bsm-secret-fetch-for-environmentfile.sh reference to
  the canonical mnemosyne_dashboard_start Hermes tool. The bsm helper is
  the ExecStartPre= BWS-secrets template, not a Mnemosyne restart helper.
  Closes E3-T17.

- docs/architecture/open-questions.md (new file): OQ-1 documents Phase 1A
  vs Phase 1C transition for AMB-Redis :6380, with live broker verification
  (redis-cli PING returns NOAUTH, requirepass confirmed, no aclfile loaded)
  and the 6 transition criteria for closing this question. Closes E3-T18.

No amend, no force-push — separate commit so remote history stays linear.