feat(roadmap): define launch success metrics baseline + 90-day targets

[mrbobbytables]

Closes #100.

The site had no defined success metrics for the KubeCon NA 2026 launch: no baseline, no targets, so an announcement post going out is unfalsifiable — there's no way to know whether it worked.

What this adds

• data/launch-metrics.json: 5 signals, each with a live-collected baseline and a target90Day: GitHub stars, GitHub watchers, GitHub forks, new unique human contributors since baseline, and good-first-issues claimed.
• scripts/collect-launch-metrics.mjs: fetches live data from the GitHub API and writes the file above, following the same generated: true / never-hand-edited pattern as data/metrics.json. Wired as npm run collect:launch-metrics.
• scripts/validate-launch-metrics.mjs: validates 3-5 signals, unique ids, required provenance (sourceUrl/collectedAt), and that baseline/target90Day are finite non-negative integers with target90Day >= baseline. Wired as npm run validate:launch-metrics, added to the read-only validator suite (tests/validators-smoke.test.mjs), and covered by tests/validate-launch-metrics.test.mjs (16 tests).
• ROADMAP.md and LAUNCH.md now point at this file as the source of truth for the launch success-metrics bullet, replacing a hand-written, already-stale table.

Why the contributor signal is a delta, not a raw count

This repo's pre-launch history already has 45 human contributors (git log/GitHub contributors API), so a raw cumulative "unique contributors" count can never sit below a small post-launch target — that was true of the number already in LAUNCH.md before this PR (baseline 1, target 5), and it doesn't hold up against the real API data. Instead, the signal is "new unique human contributors since this baseline snapshot": the file records priorContributorLogins at capture time, baseline is 0 by definition, and the 90-day re-measurement counts logins present then but absent from that list.

Why the baseline is flagged as a rehearsal capture

A previous attempt at this issue (#143) was closed as stale because it froze a months-old (2026-08-08) baseline that conflicted with the launch plan's own W-6 pre-launch-checkpoint timing (2026-09-28, per LAUNCH.md). This PR avoids repeating that: the committed data/launch-metrics.json is live data captured today, but its checkpoint field and the ROADMAP.md/LAUNCH.md prose both say explicitly to re-run npm run collect:launch-metrics at (or after) 2026-09-28 before treating the numbers as the final pre-launch baseline for the 90-day comparison.

Verification

• npm run test:unit — 501/501 passing (16 new tests in tests/validate-launch-metrics.test.mjs)
• node scripts/validate-launch-metrics.mjs — passes
• node scripts/collect-launch-metrics.mjs (with GH_TOKEN) — regenerates the committed file byte-for-byte modulo timestamps
• npm run build — succeeds
• npx prettier --check on all touched non-markdown files — passes (the touched .md files have pre-existing prettier warnings unrelated to this change)

— hive: backend=copilot model=claude-sonnet-5

🐝 Hive Agent: contributor | SHA: 0d81477

[castrojo]

Substantive review:

• Adds launch success metrics tracking (data/launch-metrics.json, scripts/collect-launch-metrics.mjs, scripts/validate-launch-metrics.mjs) for KubeCon NA 2026 launch.
• Provides test coverage in tests/validate-launch-metrics.test.mjs.
• CI is green across all checks.
• Awaiting second/independent maintainer review before merge.