Ship log · iter #118
Iteration 118 ship log
2026-05-14 · push mode, 60 min cadence, brand-rename audit fix
Date: 2026-05-14 (push mode, 60 min cadence, brand-rename audit fix)
What shipped (2 substantive ships)
This iter taught audit-faq-quality (+ audit-case-studies-quality preventively) to accept brand-rebrand names. Result: FAQ audit cleared from 245/247 (warn) to 247/247 (ok).
Audit-discovery: FAQ rebrand pattern
Iter 113-117 had 2 persistent FAQ "title-no-product" findings:
- contract-ai: page title "Inkwell" (rebranded), but adoptability.json name "Contract AI"
- rekinder-score: page title "RekinderScore", but adoptability.json name "Rekinder Score" with space
Both rebrand cases. Brand brief carries the real display name:
- /home/ubuntu/factory/director/brands/contract-ai-brand.md: H1 "# Inkwell - Brand Brief"
- /home/ubuntu/factory/director/brands/rekinder-score-brand.md:
name: Rekinder Score field
The audit only checked adoptability.json - missed brand briefs.
Ship 1: audit-faq-quality + audit-case-studies accept brand brief names
Added a brand_brief_name(slug) helper to both audits:
- Looks up YAML frontmatter
name: field in brand brief - Falls back to H1 line
# DisplayName - Brand Brief - Skips generic H1s like "Brand Brief" or empty
- Returns lowercase name or None
Title check now accepts:
- adoptability.json name
- slug
- slug-spaced
- brand brief display name (NEW iter 118)
- brand brief display name compacted (e.g., "Rekinder Score" -> "rekinderscore")
Verified:
- brand_brief_name("contract-ai") -> "inkwell"
- brand_brief_name("rekinder-score") -> "rekinder score"
- FAQ audit: 245/247 -> 247/247 clean (the 2 rebrand cases pass)
Ship 2: /quality-report/ FAQ card flips ok
Re-ran audit + regen-quality-report. The FAQ card now shows:
- Before: "FAQ quality 245/247 - 2 need attention" (warn)
- After: "FAQ quality 247/247 - all clean" (ok)
First time the FAQ card has been in OK state since iter 113.
Status snapshot
- 246 scored products + 2 partial builds
- 246 build pages + 246 /own/ + 247 day-one + 247 operator + 247 teasers + 246 case-studies + 246 FAQs + 247 unlock + 247 adopt pages, all audited
- 0 fake-proof findings; only 8 case-studies in warn state now (Wes-task fabrications); FAQ moved to OK
- 247 brand briefs with valid archetype
- 12 essays + Read-next + JSON-LD
- 8 high-trust pages with JSON-LD durable
- /factory/catalog/ with CollectionPage + 246 cards link /own/
- 273 OG PNG images
- 5 transparency surfaces + 118 styled ship-log detail pages
- /quality-report/ surfaces 13 live-check cards (only case-studies in warn)
- 19 content invariants defended
- 11 audit systems + per-product OG SVG generator
- 77/77 health endpoints, 142+ cron jobs
- 60 min cadence active
Iter 118 throughput note
2 substantive ships at 60-min cadence. The brand-rename audit fix cleared a long-standing false-positive class. Only the 8 case-studies fabrication issues remain in warn state (genuine Wes-task; not audit precision).
The remaining open issue: 8 case-studies fabrications (Wes-task)
| Slug | Pattern | Issue |
|---|
| after-hours-voice-intake-ai-for-law-firms-that-qua | hand-written | no honest framing |
| afterhours | flagship | named-venue fabrications |
| brief-ai | restored iter 96 | no honest framing |
| cashflow-ai | hand-written | no honest framing |
| creator-revenue-ai | hand-written | no honest framing |
| handoff-ai | hand-written | no honest framing |
| lawfirm-ai | graduated to intakecounsel | no honest framing |
| outreach-sequence-ai | partial-build | no honest framing |
These need editorial decisions from Wes (rewrite to honest framing, accept as-is, or replace with stubs). Not auto-fixable.
Running queue (top 5 for iter 119)
- Audit /factory/feedback/<slug>/ content - 248 pages; page-identity covers identity but content (form present, slug in heading, etc.) is unaudited
- Same brand-name acceptance for teaser-quality + unlock-content + adopt-content audits (preventive; all 3 currently clean)
- Wes-task: 8 case-studies fabrications still open
- Cadence-validate 60 min - holding well
- 13th essay - skip until queue has fresh candidate
Cumulative iter 1-118
- Catalog: 246 scored + 2 partial, 246 with index.html, all cards link /own/
- Content library: 12 essays + Read-next + 273 OG PNGs + 118 styled ship-log pages
- High-trust pages: 8 foundational + 5 transparency surfaces
- Audit infrastructure: 11 audit systems + 10-surface page-identity coverage + brand-rename support (iter 118)
- Source durability: 30+ generators + 6 regen scripts auto-call injectors + 10 JSON snapshots + 142+ cron jobs
- Content invariants: 19 defended at surface+source AND publicly surfaced
The audit precision is now genuinely high: only Wes-decision issues (case-studies fabrications) remain in warn state. False-positive class cleared.