Changelog

X Dashboard — Unified Drafts + Virality-Ranked View

The current unscheduled-drafts section is split into one table per pillar, truncates tweet bodies, shows no virality signal, and has no way to get the attached image directly. The user wants a single flat table for all drafts (no comparison “vs” posts), a separate articles section, each draft ranked by virality potential using existing SEO signals and the interview bank, the complete tweet body visible and copyable inline, and a direct image link on every row with media.

X Dashboard Single View + Article Overview with Images

Two compounding problems: the X dashboard is cluttered with sections that create noise instead of clarity, and posts are being generated as thread starters rather than complete standalone solos. Separately, the article overview is a plain table with no images and no performance signal, and articles are being published too short relative to what actually ranks.

Update Playbook §7.18 and Audit Results with Fan-Out Gaps

Task #1484 confirmed four GEO fan-out gaps from the DataForSEO “Inside 100,000 Prompts” report. Both the playbook (§7.18) and the audit results need to reflect these confirmed gaps so the documentation stays accurate and an operator reading either doc understands the current state of the system.

Fix YouTube Briefs: Never Auto-Cancel

YouTube founder briefs (owned_youtube_brief, owned_social_youtube) must never be auto-cancelled — the founder reviews and records them on their own schedule and will signal completion manually. The current 7 briefs in the DB have stale cancelled metadata from an old sweeper run (April 23), while their DB status is still awaiting_human_approval. This inconsistency makes the must-ship panel unreliable. Additionally, _build_manual_work is called twice inside build_v4_data, adding latency and a second failure point for the must-ship panel.

Close the Four GEO Fan-Out Report Gaps

The DataForSEO “Inside 100,000 Prompts” report confirmed four optimization gaps in our GEO infrastructure that are not covered by any existing task. Closing them brings our fan-out alignment fully in line with the report’s findings.

Public changelog and service status page

You’re reading it. Every Forth* product site now publishes a /changelog (with RSS and JSON feeds) and a /status link that points to our external uptime monitor’s hosted status page.

• Weekly draft proposer scans the merged shipping log and queues entries for human review at /admin/changelog
• Per-entry product visibility — entries can target one site, several sites, or all six
• Status page redirect uses the STATUS_PAGE_URL env var so the operator can swap providers without a code change

External uptime monitoring across all six product sites

We now monitor every brand homepage plus the deployed app’s /healthz endpoint from outside our infrastructure. Brand-specific body-string assertions catch the case where a generic Replit error page is served with a 200 status. Alerts fan out to email and Slack/Discord, and the public status page is wired to /status so visitors and prospects can see real-time uptime.