fleet-status: drop the retired Second Read lane

[askalf]

The Second Read has been retired: operator decision, 2026-09-25. fleet/second-read is no longer a required check on any repo, and nothing posts a verdict for it anymore, so this lane would sit at pending forever.

• scripts/fleet-status.mjs now posts only fleet/verify and fleet/review. SECOND_READ_LOGIN, secondReadAtHead and the fleet/second-read lane are gone.
• The tests drop every Second Read case. The shared checks now expect two lanes where they expected three.

The remaining fleet-status tests all pass locally with no Second Read references left in the script or its test.

[sprayberry-redline]

Automated review from the Sprayberry Labs fleet code reviewer.
Reviewed by the gating lane (gating review).

Verdict: approve. No blocking findings at head 4e74c00.

What I checked:

• The diff is a pure removal: SECOND_READ_LOGIN, secondReadAtHead, the fleet/second-read entry in CONTEXTS, and the third out.push block in laneStatuses are gone from scripts/fleet-status.mjs, and every test that depended on them is dropped from scripts/fleet-status.test.mjs. Nothing is added except a header-comment count change ("three" to "two").
• At the head, grep -i second over scripts/fleet-status.mjs, scripts/fleet-status.test.mjs, .github/workflows/fleet-status.yml, README.md and the docs tree returns nothing, so no caller of the removed exports remains.
• OWN_CONTEXTS = new Set(Object.values(CONTEXTS)) (line 22) now holds exactly fleet/verify and fleet/review, which is what the self-exclusion test own lanes required, three rounds: both green asserts. The test's own-lanes list was updated in lockstep (const own = [CONTEXTS.verify, CONTEXTS.review]).
• The a missing context is posted assertion in scripts/fleet-status.test.mjs:252 was re-pointed at a synthetic fleet/new context rather than deleted, so statusesToPost still has coverage for the missing-context branch.
• The verify and review lanes' predicates (code, verified, gated, rv.commitId === facts.head) are untouched; the removed block was a leaf that read them, not something they read.
• Required checks test, analyze (javascript-typescript) and docker-build all pass at 4e74c00.
• Tell pass over the title, body, commit message, code and test: no em dashes, no patch narration, no filler. The commit body states the fact plainly.

Minor, non-blocking: the removal leaves two consecutive blank lines before return out.map(...) in laneStatuses (scripts/fleet-status.mjs:175-176). Harmless; a linter would fold it.