docs: triage the 42 unreceipted rituals before building Fix 3

[MCKRUZ]

Proposal, not implementation. Nothing in Fix 3 is built yet — this is the triage you asked for first, and the four open questions at the end are what I need before starting.

Ships as a major version with a migration note, per your call.

The headline

Fix 3's instruction was "add an artifact spec and a registry entry for each of the 42." Having read all 42 against what they actually are, that would be wrong for 24 of them — and 42 new required files is a lot of scaffolding for a standard that warns scaffolding carrying human accountability should not grow carelessly.

	Disposition	Count
A	Required receipt + HITL gate	11
B	Optional receipt, surfaced at sign-off	14
C	Verify the real thing, don't file a document about it	6
D	No artifact — already recorded, or ceremony	6

Plus one CI mechanism and one already fixed. The arithmetic reconciles to 42 (two dispositions absorb two ledger rows each).

The test for A vs B: in an incident review, would you need to know whether this happened? If its absence would change the conclusion, it blocks.

Two findings worth your attention

Six of the eleven required artifacts already exist as optional or recommended entries. For those, Fix 3 is a promotion plus a HITL gate rather than new authorship — which materially shrinks the release.

The risk:high sign-off is not an artifact. It is currently convention in a PR comment that nothing templates and nothing checks. A markdown receipt would sit in .sdlc/ describing a merge that already happened; the check has to be at the merge. It becomes a required status check instead.

Open questions

Four, at the end of the document. The one I am least sure about is whether the PO decision record and tooling record belong in A or B — they are SOW preconditions with billing teeth, which argues for A, but duplicating contract terms into .sdlc/ risks the file and the SOW disagreeing.

Approval

Self-approved under the solo-maintainer carve-out. This is a proposal document; the work it describes has not started.

🤖 Generated with Claude Code

https://claude.ai/code/session_014aesqvsMqJpzEdpDXDTjvF

[MCKRUZ]

Closing: main already carries FIX-3-TRIAGE.md, in a newer form than this branch. Merging
this would be a revert, not an addition.

The copy on main is 200 lines; this branch's is 174. The differences are not cosmetic:

	this branch	main
Status	"proposal, awaiting Matt's review. Nothing here is implemented"	"IMPLEMENTED in claude-code-sdlc 1.0.0"
Disposition A	11 rows	11 rows → 12 files
A5	walking-skeleton-spec.md, "optional in registry"	walking-skeleton-definition.md, with the correction explaining why it is not a promotion of Phase 3's existing file
A7	one file, cold-checkout-record.md	two files, readme-verification.md + runbook-walkthrough.md
Open questions 1–4	open	resolved, with the answers recorded

So the branch predates the implementation it describes, and the four in-flight corrections that
came out of building it live only on main. The document has since done its job twice over: it
drove the 1.0.0 receipt migration, and re-reading it during the 1.0.1 work is what identified
issues #31/#33/#35 as one unfinished migration rather than three bugs.

Nothing is lost by closing — the branch stays if anyone wants the earlier draft.