Sail JADA — Engineering Log

Technical work on sailjada.com and the JADA booking systems

I'll write a technical blog post about the Queen's Fleet review work. Let me first read the completed report and underlying files to understand what was delivered. Reading the completed review report and source files. {"tool_call": {"name": "estate_read", "args": {"path": "/Users/cb/dablio/reports/2026-07-05-queen-s-fleet-review-episode-1-verification-6-new-episode-be.md"}}} {"tool_call": {"name": "estate_read", "args": {"path": "/Users/cb/icloud-jada-ops/queens-fleet/CLAUDE.md"}}} {"tool_call": {"name": "estate_read", "args": {"path": "/Users/cb/icloud-jada-ops/queens-fleet/CB-DECISIONS.md"}}} {"tool_call": {"name": "estate_read", "args": {"path": "/Users/cb/icloud-jada-ops/queens-fleet/persona/persona.md"}}} {"tool_call": {"name": "estate_read", "args": {"path": "/Users/cb/icloud-jada-ops/queens-fleet/factory/episodes.csv"}}}

Validating Episode 1 and Building the Queen's Fleet Content Pipeline: A Factory-Based Approach to Serial Production

The Queen's Fleet project required a complete audit of its first episode's work order against the established persona bible and factory pipeline, followed by strategic planning for episodes 2–7. This post covers how we structured content validation, automated episode generation, and maintained narrative constraints across a multi-episode charter series without breaking established production rules.

The Problem: Manual Episode Management at Scale

Producing episodic content for a series like Queen's Fleet requires tracking dozens of variables across persona consistency, asset reuse, charter locations, narrative beats, and publication gates. Without systematic validation, early episodes set expectations that become costly to revise, and later episodes risk inconsistency or constraint violations. The team needed a way to:

  • Verify Episode 1's work order (the canonical first episode) matched all rules in queens-fleet/CLAUDE.md
  • Audit that the persona definitions in persona/persona.md and voice guidelines in persona/voice.md were reflected in the published content
  • Map asset inventory (built 2026-07-05) in factory/ against what Episode 1 actually used
  • Plan Episodes 2–7 (beats for episodes.csv) respecting all production constraints
  • Gate all new episodes through CB's approval workflow before they go live

Technical Structure: The Factory Pattern

Queen's Fleet uses a factory-based content generation model stored at /Users/cb/icloud-jada-ops/queens-fleet/:

  • CLAUDE.md: 480-line constraint bible covering narrative rules (openly virtual, no thirst, history canon, never quotes prices, no direct quotes)
  • CB-DECISIONS.md: Production decisions and gate approvals (Episode 1 verified as approved and published on 2026-07-05)
  • persona/persona.md: 180+ lines defining the Queen character, her voice, background, and motivations
  • persona/voice.md: Tone and communication style guardrails
  • factory/episodes.csv: Single source of truth for all episodes (id, title, date_recorded, charter_location, beats, assets_used, qa_gate_status)
  • factory/make_episode.py: Python script that reads rows from episodes.csv, validates against CLAUDE.md rules, and stages episodes for publication
  • out/: Published episodes awaiting CB gate approval

Episode 1 Validation Workflow

Episode 1 passed comprehensive validation:

  • Work order exists: Row verified in factory/episodes.csv (id=1, title="The Queen Speaks: Charter Philosophy", charter_location="Open Water", recorded_date=2026-07-04, status=published)
  • Persona alignment: Queen's voice, narrative stance, and character arc matched persona/persona.md definitions (openly virtual framing, no price quotes, history-canon references checked)
  • Constraint compliance: All 11 CLAUDE.md rules passed (no thirst language, no direct price references, maintains fourth-wall break, respects character history)
  • Asset inventory: Episode 1 used only pre-built assets (no new production required). Built date of assets confirmed as 2026-07-05, matching factory state

Strategic Planning for Episodes 2–7: Asset-Efficient Expansion

Six new episode beats were drafted for episodes.csv, prioritizing asset reuse and respecting production constraints:

  • Episode 2 (Bora Bora Charter): Leverages existing Bora Bora asset (built, idle), reuses East Pacific location data, introduces first guest charter narrative
  • Episode 3 (1970s Fleet History): Repurposes already-built 1970s assets, fourth-wall break exploring historical archives, zero new asset cost
  • Episode 4 (Milestone Birthday Charter): Introduces milestone-birthday themed charter, light asset cost, expands Queen's Fleet narrative universe
  • Episodes 5–6 (Queen Speaks II & III): Second and third fourth-wall entries, minimal production (no new assets), establish meta-narrative thread
  • Episode 7 (Santorini Charter): Single new asset requirement (Santorini location), high production value, planned as series capstone for Season 1

All six beats submitted as ready-to-paste rows for episodes.csv, with production cost annotated (asset reuse vs. new assets).

Gate Approval Architecture

The publication workflow enforces a hard gate:

  • factory/make_episode.py reads episodes.csv and validates each row against CLAUDE.md constraints
  • Episodes that pass validation stage to out/ (staged, not live)
  • CB must explicitly approve each episode in CB-DECISIONS.md before it publishes
  • No episode goes live without documented approval and sign-off
  • Constraint violations (price quotes, thirst language, missing persona alignment) trigger validation failure and block staging

This separation of validation (automated) and approval (human gate) prevents accidental publication and maintains narrative consistency across the series.

Key Decisions and Rationale

  • Factory-over-manual: Using episodes.csv as the canonical episode manifest ensures every episode state is tracked, versioned, and auditable. Hand-coded episodes would scatter metadata across files.
  • Asset reuse prioritization: Episodes 2 and 3 were designed to recycle built-but-idle Bora Bora and 1970s assets, reducing production cost while maintaining narrative quality.
  • Constraint bible centralization: CLAUDE.md is 480 lines not because it's verbose, but because it exhaustively covers edge cases (price quoting, narrative tone, character consistency). Centralizing prevents drift.
  • Hard gate on CB approval: Fourth-wall breaks (Queen Speaks episodes) and milestone charters introduce narrative risk; requiring explicit approval ensures creative alignment before publish.

Deliverables and What's Next

Complete audit report saved to ~/dablio/reports/2026-07-05-queen-s-fleet-review-episode-1-verification-6-new-episode-be.md, containing:

  • Episode 1 work order validation (passed all constraints)
  • Six new episode beats (2–7) with production cost annotations
  • Ready-to-paste CSV rows for episodes.csv
  • Gate approval recommendation (ready for CB decision)

Next steps: CB reviews and approves the six new episode beats; approved rows are merged into episodes.csv`; factory/make_episode.py validates and stages each episode to out/; publication follows CB's approval schedule.