Skip to content

docs: the M13 record — a home-screen shortcut that stays logged in, and the roadmap row (#239) - #243

Merged
davison merged 4 commits into
mainfrom
docs/m13-record
Sep 30, 2026
Merged

davison merged 4 commits into
mainfrom
docs/m13-record

Conversation

@davison

@davison davison commented Sep 30, 2026 •

Copy link
Copy Markdown
Owner

The milestone record for M13, milestone issue #239, on the housekeeping path (SPEC §4): no task, no plan. gh codecrew milestone close 13 closes the milestone, not this merge.

What it adds

  • docs/milestones/13-a-home-screen-shortcut-that-stays-logged-in.md, covering:

    • the goal and outcome, from capture The e-ink tablet's home-screen shortcut asks for the token every time: SameSite=Strict withholds the cookie #238 and the operator's device reports;
    • the SameSite=Lax decision: the operator's, taken before the milestone opened, with his probe of a Lax build, the trade-off, what it supersedes (Tailnet access: extra Host name, token login and session cookie #39's Strict and M12-R4), and the rejected install-the-app alternative;
    • the refusal log: reasons, the previous-token check, escaping, every refusal logged, and the bounds, with review round one (log forging, unlogged refusals) and round two (the unbounded wrong-token line);
    • the GET audit;
    • GitHub's keyword-linking regression (community discussions 209162 and 209148), the manual link, and the commit-body keyword through which task 240 was shut;
    • requirement outcomes (R1 to R3 satisfied, R4 untestable until this merges), QA's findings and their disposition, including the QA process slip, and the captures;
    • the release: v0.3.2, still to be tagged.
  • ROADMAP.md: the M13 row, Done.

  • Front door:

    • docs/introduction.md: the opening ends at milestone thirteen, with an M13 paragraph; What holds these numbers counts 124 checks and names the tailnet suite's cross-site cases.
    • CONTRIBUTING.md: the make e2e paragraph names the cross-site cases and counts 124 tests.

    The README makes no claim M13 changed.

Gaps the record names

Checks

make check passes. gh codecrew milestone evidence 13 reports that all cited links resolve, and every github.com URL in the record was checked by hand as well.

Refs #239

🤖 Generated with Claude Code

davison and others added 2 commits September 30, 2026 16:48
…nd the roadmap row (#239)

The milestone record on the housekeeping path (SPEC §4): the operator's
Lax decision with its probe, trade-off and what it supersedes, the
refusal log and the two review rounds that bounded and escaped it, the
GET audit, the GitHub linking regression and the commit-body workaround,
the requirement outcomes, the captures, and v0.3.2 still to be tagged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… browser checks (#239)

Part of the milestone record on the housekeeping path (SPEC §4): the
introduction's opening now ends at milestone thirteen and says what it
changed, and the introduction and CONTRIBUTING count 124 e2e tests and
name the tailnet suite's cross-site checks.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…239)

Part of the milestone record on the housekeeping path (SPEC §4): the
manual link, the 209148 workaround, the auto-close check, the search ask
and the operator confirmation now cite the coordinator's attribution
Decision on #239, still marked as recorded late with the record's
Deviation cited; the e-ink guide's NeoBrowser claims are captured as #244.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@davison

davison commented Sep 30, 2026

Copy link
Copy Markdown
Owner Author

Review — 6b16c84

Reviewer seat (~, operator auth), clean context, in a detached worktree at 6b16c84. This was the head I was given, and it is still the PR head. gh codecrew version reports protocol 2.1, which matches the floor in .codecrew/config.yml. I read the diff first. Then I read the charter: #239's body and all 6 comments (scope, QA verdicts, observations, disposition, the record's Deviation, attribution). Then #240 (plan, both Decisions, the correction, QA's finding, the full timeline), PR #241 (body, the three review rounds, the confirmation, its timeline), captures #238 (with the probe), #242 and #244, the #39 cookie Decision, upstream radiusred/gh-codecrew#386, and community discussions 209162 and 209148 with all their comments. I read the PR body last. I held the record to the doc-synthesizer contract, including this hub's front-door list, and used the M12 record and its review as the bar.

Findings

1. nit: 83d0efc is described as a docs commit. docs/milestones/13-a-home-screen-shortcut-that-stays-logged-in.md:278-280 says the fix pass added "83d0efc and 45bb114 (the docs for nits 4 to 6 and finding 2)". 83d0efc is test(session): gofmt the reason table, on main as da138d8. Its whole diff is internal/session/session_test.go | 2 +- (git show --stat da138d8), and round two says it is "session_test.go only, reordering one import line". The record says the same thing itself at :377-379, so the two passages contradict each other. The docs were 45bb114 alone. Nit 6's code change (the "in the minute from" wording) went in with b107985 (git show c887cdc | grep 'in the minute from' finds it in refusals.go's caller and the test). Fix: "45bb114 (the docs for nits 4 to 6 and finding 2), and 83d0efc (gofmt of a test)".

2. nit: the reason for fix(man) is not on the trail the record cites. :65-67 says "The man page took fix(man), so that CI runs for it". The cited scope Decision says nothing about commit types, and the plan says only "fix(man) if the man page changes", with no reason. The reason is true, but it comes from this hub's reviewer extension (#215), and round one's check was "correctly not docs" (PR #241). Also, "none is feat, as the plan and the release Decision expected" cites only the scope Decision. That Decision says "The change is a fix, so it is a patch". It is the plan that says "No feat". Fix: cite #240's plan for "no feat" and round one for fix(man). Or drop "so that CI runs for it".

3. nit: "Three of the operator's words" lists four. :469-477 heads the list "Three of the operator's words were recorded late". Its second bullet also carries the search ask ("and the search ask that found the two discussions"). That is a fourth relay. The Deviation (5914650687) did not name it, and only the attribution Decision records it. :479-483 then says the Deviation "named the gap" and the Decision "records all of them", which reads as though the Deviation named the search ask too. The PR body repeats "Three". Fix: "Four", or keep three and add one sentence saying the attribution Decision also recorded the search ask, which the Deviation had not named.

Nothing blocking.

What I verified, and how

Capture candidates

None.

Verdict: approve

Part of the milestone record on the housekeeping path (SPEC §4): 83d0efc
is named as the test(session) commit it is; "no feat" cites #240's plan
and review round one, and the unsourced reason for fix(man) is dropped;
the late-recorded words count four, with the Deviation naming three and
the attribution Decision the search ask as well.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@davison

davison commented Sep 30, 2026

Copy link
Copy Markdown
Owner Author

Operator confirmation: this record, at head 628ef53, is accepted for the operator (pure solo tier, SPEC §4). The acceptance rests on his standing merge confirmation for this milestone and its record ("release v0.3.2"), recorded in #239 (comment) and relayed from the coordinator session.

The model review approved at 6b16c84 (#243 (comment)). Commit 628ef53 takes all three nits, which are wording only. The PR body was also reworded so that it closes nothing (closingIssuesReferences: []). The gate passed; check and e2e were skipped because every commit is docs. The doc-synthesizer may merge.

@davison
davison merged commit f417356 into main Sep 30, 2026
3 checks passed
@davison
davison deleted the docs/m13-record branch September 30, 2026 16:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant