Commit Graph

2 Commits

Author SHA1 Message Date
8ddcf0dd27 Brainstorm Q8: greenfield-mode (fresh-enterprise bootstrap)
Surfaced 2026-05-06 during factory-bootstrap field-test on Win11 laptop.
Decision: factory MUST support installing a fresh empty enterprise
(no existing gitea content) — not just cloning the existing one.

Three of five setup-* skills already cover greenfield (setup-wiki,
setup-tasks, project-bootstrap). Two need init-mode (projects-meta,
interns). Six root meta-repos have no templates anywhere — must add
templates/enterprise/<repo>/ skeletons. Anti-rec: do NOT bootstrap
the git-host itself — that's L0a infra, not L1 (factory).

Implementation deferred to post-v0; not blocking clone-mode v0.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-06 15:52:43 +03:00
f00e2c8f9e Brainstorm buffer: factory-bootstrap (in progress)
Design buffer for the software-factory bootstrap discussion. Captures:
- diagnosis of current dot-folder infrastructure (.common, .organization,
  .templates, .wiki, .tasks, .meeting-room, .factory)
- decision: factory lives at ~/projects/.factory/ (option a, sibling of
  .common), gitea repo named 'factory' without leading dot
- L0a/L0b/L1/L2 architecture
- closed Q1 (Go for L1 binary), Q6 (mise + native pkg-mgr for toolchain)
- deferred Q7 (access control to v2)
- live field-test on a fresh Win11 laptop, log in
  ../.factory/L0/install-log.md (gitea repo: factory)
- resume point recorded; user paused mid-step 11b

Buffer remains active — promotion to shared wiki happens via
meeting-room-promote-brainstorm once Q2-Q5 are also closed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-06 11:55:33 +03:00