@growthub/cli 0.14.5 → 0.14.8

package/assets/worker-kits/growthub-open-montage-studio-v1/skills.md

@@ -1,254 +0,0 @@
-# Open Montage Studio — Master Skill Document
-
-This is the canonical methodology reference for the Open Montage Studio Operator. Read this file at the start of every session before producing any artifacts.
-
----
-
-## 1. Workflow Order
-
-The operator follows a strict 10-step workflow defined in `workers/open-montage-studio-operator/CLAUDE.md`. This skill doc provides the detailed methodology for steps 5 through 9.
-
----
-
-## 2. Pipeline Selection Logic
-
-OpenMontage supports 12 production pipelines. Selection is a structured decision, not a guess.
-
-### Decision Tree
-
-```
-Intent: Explain a topic to an audience?
-  YES → Animated Explainer
-
-Intent: Create motion graphics, kinetic typography, or abstract animation?
-  YES → Animation
-
-Intent: Avatar-driven presenter video?
-  YES → Avatar Spokesperson
-
-Intent: Cinematic trailer, teaser, or mood-driven brand film?
-  YES → Cinematic
-
-Intent: Batch of ranked short-form clips from one long source?
-  YES → Clip Factory
-
-Intent: Thematic montage from real stock footage or open archives?
-  YES → Documentary Montage
-
-Intent: Enhance existing footage with AI-generated support visuals?
-  YES → Hybrid
-
-Intent: Subtitle, dub, or translate existing video into other languages?
-  YES → Localization & Dub
-
-Intent: Turn podcast highlights into video?
-  YES → Podcast Repurpose
-
-Intent: Polished software screen recording or walkthrough?
-  YES → Screen Demo
-
-Intent: Footage-led speaker video, presentation, or vlog?
-  YES → Talking Head
-```
-
-If the intent spans two pipelines, pick the primary and document the secondary as a composition layer or follow-up pipeline.
-
-### Pipeline Stage Flow (Universal)
-
-Every pipeline follows the same structured flow:
-
-```
-research → proposal → script → scene_plan → assets → edit → compose
-```
-
-Each stage has a dedicated director skill in the OpenMontage fork at `skills/pipelines/<pipeline-name>/`.
-
----
-
-## 3. Provider Scoring Methodology
-
-Every tool selection runs through a 7-dimension scoring engine.
-
-| Dimension | Weight | What It Measures |
-|-----------|--------|------------------|
-| Task fit | 30% | How well the tool matches the specific production need |
-| Output quality | 20% | Resolution, fidelity, consistency of results |
-| Control features | 15% | Prompt adherence, style control, parameter granularity |
-| Reliability | 15% | Uptime, error rate, queue latency |
-| Cost efficiency | 10% | Price per output unit relative to quality |
-| Latency | 5% | Time from request to deliverable |
-| Continuity | 5% | Session persistence, resumability, rate limits |
-
-### Provider Tiers
-
-**Zero-cost (always available):**
-- Piper TTS — offline narration
-- Archive.org + NASA + Wikimedia — free archival footage
-- Remotion — React-based programmatic video composition
-- FFmpeg — encoding, subtitles, audio mixing, color grading
-
-**Free-key (developer accounts, no charge):**
-- Pexels, Pixabay, Unsplash — stock media
-- Google TTS — 1M chars/month free
-- ElevenLabs — 10K chars/month free
-
-**Pay-as-you-go:**
-- fal.ai (FAL_KEY) — FLUX images + Kling/Veo/MiniMax video
-- OpenAI — DALL-E 3 + TTS
-- Google Imagen — $0.02-0.06/image
-- xAI Grok — images + video
-- Runway Gen-4 — highest quality AI video
-- Higgsfield — multi-model video orchestrator
-- HeyGen — avatar videos
-- Suno — AI music generation
-
-**Local GPU (free, requires hardware):**
-- WAN 2.1, Hunyuan, CogVideo, LTX-Video — local video generation
-- Stable Diffusion — local image generation
-
----
-
-## 4. CMS Node Bridge Mapping
-
-GrowthHub CMS nodes expose video and image generation as hosted execution capabilities. The bridge maps these into the OpenMontage production flow.
-
-### How It Works
-
-1. Agent queries the CMS capability registry for available nodes in the `video` and `image` families
-2. CMS nodes execute via hosted execution client, producing output URLs
-3. Output URLs are treated as source assets in the OpenMontage pipeline
-4. OpenMontage handles composition, post-production, and final rendering
-
-### CMS Family → OpenMontage Mapping
-
-| CMS Node Family | CMS Execution Kind | OpenMontage Integration Point |
-|-----------------|--------------------|-----------------------------|
-| `video` | `hosted-execute` | Source clips for `tools/video/` input or direct composition via Remotion/FFmpeg |
-| `image` | `hosted-execute` | Scene images for `tools/graphics/` input or Remotion animated composition |
-| `slides` | `hosted-execute` | Slide deck assets for Screen Demo pipeline |
-| `text` | `hosted-execute` | Script/narration text for TTS processing |
-
-### CMS Node Output → OpenMontage Asset Type
-
-| CMS Output Type | OpenMontage Asset Category | Typical Usage |
-|-----------------|---------------------------|---------------|
-| Video URL | `source_clip` | Scene footage, B-roll, transition clips |
-| Image URL | `scene_image` | Hero shots, backgrounds, product stills |
-| Audio URL | `narration` or `music` | Voiceover, soundtrack |
-| Text | `script` | Script content for TTS conversion |
-
-### When to Use CMS Nodes vs Direct Providers
-
-Use CMS nodes when:
-- The user has an active GrowthHub session
-- The generation type matches an available CMS node
-- The user wants unified billing through GrowthHub
-- The user does not have direct provider API keys
-
-Use direct providers when:
-- Local-fork mode with full tool access
-- The specific tool/model is not available as a CMS node
-- Lower latency is needed (direct API vs hosted execution)
-- The user has provider keys and prefers direct access
-
----
-
-## 5. Prompt Planning Rules
-
-### Image Prompts
-
-1. Lead with the subject and composition
-2. Include style directive (photorealistic, illustrated, anime, etc.)
-3. Specify lighting and mood
-4. Add technical parameters (aspect ratio, resolution)
-5. Include negative prompts when the provider supports them
-
-### Video Prompts
-
-1. Describe the motion and camera movement
-2. Specify duration target
-3. Reference the source image if using image-to-video
-4. Include style consistency directives
-5. Note any audio sync requirements
-
-### Scene Sequencing
-
-1. Each scene maps to one pipeline stage asset
-2. Scenes must have explicit transition types (cut, crossfade, wipe, zoom)
-3. Audio cues align to scene boundaries
-4. Subtitle timing derives from script word timestamps
-
----
-
-## 6. Output Artifact Order
-
-Produce artifacts in this strict order:
-
-1. **VideoProductionBrief** — What we are making, why, for whom
-2. **PipelineSelectionBrief** — Which pipeline, why, fallback
-3. **ProviderSelectionBrief** — Which providers/CMS nodes, scoring, cost estimate
-4. **ScenePlan** — Scene-by-scene breakdown with timing, transitions, audio cues
-5. **PromptMatrix** — Per-scene prompts for all generation tools
-6. **GenerationBatchPlan** — Execution order, dependencies, parallel opportunities
-7. **AssetTracking** — Generated asset inventory with URLs, statuses, costs
-8. **ReviewQAChecklist** — Quality gates, validation points, approval criteria
-9. **PlatformReadyExecutionHandoff** — Complete handoff doc for execution
-
----
-
-## 7. QA Standards
-
-### Pre-Compose Validation
-
-Before rendering:
-- All scene assets exist and are accessible
-- Audio tracks are present and duration-matched
-- Subtitle timing is aligned
-- Delivery promise is achievable with available tools
-- Budget is within approved cap
-
-### Post-Render Self-Review
-
-After rendering:
-- ffprobe validation (codec, resolution, duration, bitrate)
-- Frame extraction at key points (first, middle, last, transitions)
-- Audio level analysis (no clipping, consistent volume)
-- Subtitle readability check
-- Delivery promise verification
-
-### Slideshow Risk Check
-
-If the output is image-based video, score across 6 dimensions:
-- Motion variety (camera moves, parallax, Ken Burns)
-- Transition diversity (not all cuts)
-- Animation layer count (particles, overlays, text motion)
-- Audio-visual sync
-- Pacing variation
-- Visual grammar complexity
-
-A score below 60% triggers a warning and requires explicit user approval to proceed.
-
----
-
-## 8. Budget Governance
-
-- Estimate costs before execution
-- Reserve budget before each provider call
-- Reconcile actual spend after each call
-- Default per-action approval threshold: $0.50
-- Default total budget cap: $10
-- CMS node costs flow through GrowthHub billing (no separate tracking needed)
-- Always present cost summary before starting generation
-
----
-
-## 9. Handoff Standards
-
-Every handoff document must include:
-- Execution mode (local-fork / agent-only / hybrid)
-- Complete asset manifest with URLs or file paths
-- Provider/CMS node execution instructions
-- Post-production steps (composition, audio, subtitles, color grade)
-- Render profile (resolution, aspect ratio, codec, target platform)
-- Estimated total cost and actual spend
-- Known limitations or manual steps required

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/asset-tracking.md

@@ -1,46 +0,0 @@
-# Asset Tracking
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Date | `YYYY-MM-DD` |
-
-## Generated Assets
-
-| ID | Scene | Type | Provider / CMS Node | Status | URL / Path | Cost | Notes |
-|---|---|---|---|---|---|---|---|
-| A-001 | 1 | image | | `<pending / generating / ready / failed>` | | | |
-| A-002 | 2 | image | | | | | |
-| A-003 | 1 | video | | | | | |
-| A-004 | — | narration | | | | | |
-| A-005 | — | music | | | | | |
-
-## Reuse Inventory
-
-<!-- Assets from previous sessions or existing brand kit that can be reused. -->
-
-| Asset | Source | Reusable For | Notes |
-|---|---|---|---|
-| | | | |
-
-## CMS Node Outputs
-
-<!-- Track assets generated through GrowthHub CMS nodes separately for billing clarity. -->
-
-| CMS Node Slug | Execution ID | Output Type | URL | Status |
-|---|---|---|---|---|
-| | | | | |
-
-## Cost Ledger
-
-| Category | Budgeted | Actual | Delta |
-|---|---|---|---|
-| Images | | | |
-| Video | | | |
-| Audio | | | |
-| Music | | | |
-| CMS nodes | — | — | — |
-| **Total** | **$—** | **$—** | **$—** |

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/cms-node-pipeline-mapping.md

@@ -1,64 +0,0 @@
-# CMS Node → OpenMontage Pipeline Mapping
-
-## Purpose
-
-This template documents how GrowthHub CMS node outputs integrate into the selected OpenMontage pipeline. Fill this out when the production plan uses CMS nodes for any generation step.
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Pipeline | `<selected-pipeline>` |
-| Date | `YYYY-MM-DD` |
-
-## CMS Nodes in Use
-
-| CMS Node Slug | Family | Execution Kind | Role in Pipeline |
-|---|---|---|---|
-| | `video` / `image` / `text` | `hosted-execute` | |
-
-## Mapping: CMS Output → OpenMontage Input
-
-| CMS Node | Output Type | Output Format | OpenMontage Stage | OpenMontage Tool Category | Integration Method |
-|---|---|---|---|---|---|
-| | Video URL | mp4 | `assets` | `tools/video/` | Download URL → source clip |
-| | Image URL | png/jpg | `assets` | `tools/graphics/` | Download URL → scene image |
-| | Audio URL | mp3/wav | `edit` | `tools/audio/` | Download URL → narration track |
-| | Text | string | `script` | — | Direct script content |
-
-## Execution Flow
-
-```
-1. Agent identifies required CMS nodes from the scene plan
-2. Agent prepares input bindings for each CMS node
-3. CMS nodes execute via hosted execution client
-4. Agent collects output URLs from execution results
-5. Output URLs feed into OpenMontage pipeline at the appropriate stage
-6. OpenMontage handles composition and post-production
-```
-
-## Input Bindings Per Node
-
-### CMS Node: `<slug>`
-
-| Input Key | Type | Value | Source |
-|---|---|---|---|
-| | | | `<from scene plan / user input / previous node output>` |
-
-## Output Routing
-
-### CMS Node: `<slug>`
-
-| Output Key | Type | Routes To | OpenMontage Asset ID |
-|---|---|---|---|
-| | | Scene `<N>` source clip / image | `A-<NNN>` |
-
-## Fallback Plan
-
-If a CMS node execution fails:
-
-| CMS Node | Fallback Provider | Fallback Tool | Cost Impact |
-|---|---|---|---|
-| | | | |

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/generation-batch-plan.md

@@ -1,70 +0,0 @@
-# Generation Batch Plan
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Pipeline | `<selected-pipeline>` |
-| Date | `YYYY-MM-DD` |
-
-## Execution Order
-
-Batches run top-to-bottom. Items within a batch can run in parallel.
-
-### Batch 1 — Research & Script
-
-| Step | Tool / Action | Input | Output | Depends On |
-|---|---|---|---|---|
-| 1.1 | Web search | Topic keywords | Research brief | — |
-| 1.2 | Script draft | Research brief | Script | 1.1 |
-
-### Batch 2 — Image Generation
-
-| Step | Provider / CMS Node | Prompt Ref | Input Assets | Expected Output | Est. Cost |
-|---|---|---|---|---|---|
-| 2.1 | | Scene 1 prompt | — | Scene 1 image | |
-| 2.2 | | Scene 2 prompt | — | Scene 2 image | |
-| 2.3 | | Scene 3 prompt | — | Scene 3 image | |
-
-### Batch 3 — Video Generation (if applicable)
-
-| Step | Provider / CMS Node | Prompt Ref | Source Image | Duration | Est. Cost |
-|---|---|---|---|---|---|
-| 3.1 | | Scene 1 video prompt | Scene 1 image | 5s | |
-| 3.2 | | Scene 2 video prompt | Scene 2 image | 5s | |
-
-### Batch 4 — Audio
-
-| Step | Provider | Type | Input | Duration | Est. Cost |
-|---|---|---|---|---|---|
-| 4.1 | | TTS narration | Script | — | |
-| 4.2 | | Background music | Genre/mood spec | — | |
-
-### Batch 5 — Composition & Post-Production
-
-| Step | Tool | Input | Output | Depends On |
-|---|---|---|---|---|
-| 5.1 | Remotion / FFmpeg | All scene assets + audio | Draft render | Batches 2-4 |
-| 5.2 | Subtitle gen | TTS audio | SRT/VTT | 4.1 |
-| 5.3 | FFmpeg | Draft render + subtitles | Final render | 5.1, 5.2 |
-
-## Cost Summary
-
-| Batch | Estimated Cost |
-|---|---|
-| Batch 1 (Research) | $0.00 |
-| Batch 2 (Images) | $— |
-| Batch 3 (Video) | $— |
-| Batch 4 (Audio) | $— |
-| Batch 5 (Post) | $0.00 |
-| **Total** | **$—** |
-
-## Budget Gate
-
-| Field | Value |
-|---|---|
-| Budget cap | $10.00 |
-| Estimated total | $— |
-| Approved | `<yes / pending>` |

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/pipeline-selection-brief.md

@@ -1,67 +0,0 @@
-# Pipeline Selection Brief
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Date | `YYYY-MM-DD` |
-
-## Requested Outcome
-
-<!-- What the user asked for in their own words. -->
-
-## Selected Pipeline
-
-| Field | Value |
-|---|---|
-| Pipeline | `<Animated Explainer / Animation / Avatar Spokesperson / Cinematic / Clip Factory / Documentary Montage / Hybrid / Localization & Dub / Podcast Repurpose / Screen Demo / Talking Head>` |
-| Reason | <!-- Why this pipeline is the best match --> |
-
-## Pipeline Fit Assessment
-
-| Criterion | Score (1-5) | Notes |
-|---|---|---|
-| Content type match | | |
-| Available tool coverage | | |
-| Source asset alignment | | |
-| Budget feasibility | | |
-| Quality ceiling | | |
-
-## Input Asset Requirements
-
-<!-- What the selected pipeline needs to execute. -->
-
-## Output Specification
-
-| Field | Value |
-|---|---|
-| Primary deliverable | |
-| Duration | |
-| Resolution | |
-| Codec | |
-
-## Fallback Pipeline
-
-| Field | Value |
-|---|---|
-| Pipeline | `<fallback-pipeline>` |
-| Trigger | <!-- When to switch --> |
-| Trade-off | <!-- What changes with the fallback --> |
-
-## Pipeline Stage Plan
-
-```
-research → proposal → script → scene_plan → assets → edit → compose
-```
-
-| Stage | Owner | Tool / CMS Node | Notes |
-|---|---|---|---|
-| Research | Agent | Web search | |
-| Proposal | Agent | — | |
-| Script | Agent | — | |
-| Scene plan | Agent | — | |
-| Assets | Agent + Provider | `<tool-or-cms-node>` | |
-| Edit | Agent | FFmpeg / Remotion | |
-| Compose | Agent | Remotion / FFmpeg | |

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/platform-ready-execution-handoff.md

@@ -1,103 +0,0 @@
-# Platform-Ready Execution Handoff
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Pipeline | `<selected-pipeline>` |
-| Execution mode | `<local-fork / agent-only / hybrid>` |
-| Date | `YYYY-MM-DD` |
-
-## Deliverable Summary
-
-| Field | Value |
-|---|---|
-| Video title | |
-| Duration | `<seconds>` |
-| Resolution | `<WxH>` |
-| Aspect ratio | `<ratio>` |
-| Platform target | `<YouTube / TikTok / Instagram / LinkedIn>` |
-| Estimated cost | `$—` |
-| Actual cost | `$—` |
-
-## Execution Instructions
-
-### For local-fork mode
-
-```bash
-cd ~/OpenMontage
-
-# 1. Verify environment
-python -c "from tools.tool_registry import registry; registry.discover(); print('OK')"
-
-# 2. Run the pipeline
-# The agent executes each stage in order:
-#   research → proposal → script → scene_plan → assets → edit → compose
-
-# 3. Final render location
-# Output: output/<project-slug>/final-render.<ext>
-```
-
-### For agent-only mode (CMS node generation)
-
-1. Execute CMS nodes via GrowthHub hosted execution:
-
-| Step | CMS Node Slug | Inputs | Expected Output |
-|---|---|---|---|
-| 1 | | | |
-| 2 | | | |
-
-2. Collect output URLs from CMS node execution results
-3. Follow post-production guidance below
-
-### For hybrid mode
-
-1. Generate assets via CMS nodes (see agent-only instructions above)
-2. Run local OpenMontage post-production:
-
-```bash
-cd ~/OpenMontage
-# Composition, audio mixing, subtitle burn-in, color grading
-# Use the scene plan and prompt matrix from this handoff
-```
-
-## Asset Manifest
-
-| ID | Type | Source | URL / Path | Status |
-|---|---|---|---|---|
-| | | | | |
-
-## Post-Production Steps
-
-| Step | Tool | Input | Output | Notes |
-|---|---|---|---|---|
-| 1 | Remotion / FFmpeg | Scene assets | Draft composition | |
-| 2 | Subtitle gen | Narration audio | SRT/VTT file | |
-| 3 | FFmpeg | Draft + subtitles | Subtitled render | |
-| 4 | FFmpeg | Subtitled render | Color-graded final | Optional |
-
-## Render Profile
-
-| Field | Value |
-|---|---|
-| Resolution | |
-| FPS | |
-| Codec | |
-| Bitrate | |
-| Audio codec | |
-| Container | |
-
-## Known Limitations
-
-<!-- Any manual steps, unsupported features, or quality compromises. -->
-
-## Approval
-
-| Gate | Status |
-|---|---|
-| Production brief approved | `<yes / no>` |
-| Budget approved | `<yes / no>` |
-| QA checklist passed | `<yes / no>` |
-| Ready for execution | `<yes / no>` |

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/project.md

@@ -1,48 +0,0 @@
----
-# .growthub-fork/project.md — session memory primitive (v1)
-#
-# Written by the CLI at init/import time. Append-only, human-readable; sits
-# alongside the machine-readable `trace.jsonl`. Approvals, self-eval outcomes,
-# and sub-skill runs all record here.
-#
-# Replaceable tokens (seeded at init time):
-#   {{KIT_ID}}   {{FORK_ID}}   {{STARTED_AT}}   {{SOURCE}}   {{SOURCE_REF}}
-
-kitId: "{{KIT_ID}}"
-forkId: "{{FORK_ID}}"
-startedAt: "{{STARTED_AT}}"
-source: "{{SOURCE}}"
-sourceRef: "{{SOURCE_REF}}"
-skillManifestVersion: 1
-approvals: []
-selfEvalHistory: []
-subSkillRuns: []
----
-
-# Project journal — fork `{{FORK_ID}}`
-
-Baseline primitive: `{{KIT_ID}}`. Seeded at `{{STARTED_AT}}` from `{{SOURCE}}` (`{{SOURCE_REF}}`).
-
-This is your cross-session continuity surface. Read the last few entries before every session; append after every material change, approval, and self-eval boundary.
-
-## How to use this file
-
-1. **Session start.** Read the last 3 entries + tail 20 events in `.growthub-fork/trace.jsonl`. State your plan.
-2. **Material change.** Record as a dated bullet (what, why, outcome). Simultaneously append a typed event to `trace.jsonl` via the Fork Sync Agent — never bypass.
-3. **Approval boundary.** Add a row to `approvals` in the frontmatter (verbatim decision).
-4. **Self-eval boundary.** Append to `selfEvalHistory`: `attempt`, `criteria`, `outcome`, `notes`. Enforce `maxRetries: 3`.
-5. **Sub-skill spawn.** Append to `subSkillRuns`: `subSkill`, `startedAt`, `result`.
-
-## Session log
-
-<!-- Append newest entries at the bottom. Format:
-
-### {{YYYY-MM-DD HH:mm UTC}} · {{skill-slug}}
-- **Plan.** What you intend to do.
-- **Changes.** Material changes (each must correspond to a trace event).
-- **Outcome.** Pass / fail / parked. Reference self-eval criteria.
-- **Next.** What the next session should pick up.
-
--->
-
-_No sessions recorded yet. Populated as the fork is operated._

package/assets/worker-kits/growthub-open-montage-studio-v1/templates/prompt-matrix.md

@@ -1,48 +0,0 @@
-# Prompt Matrix
-
-## Project Reference
-
-| Field | Value |
-|---|---|
-| Client | `<client-name>` |
-| Project | `<project-name>` |
-| Pipeline | `<selected-pipeline>` |
-| Date | `YYYY-MM-DD` |
-
-## Image Generation Prompts
-
-| Scene | Provider / CMS Node | Model | Prompt | Style Directive | Aspect Ratio | Negative Prompt | Expected Cost |
-|---|---|---|---|---|---|---|---|
-| 1 | | | | | | | |
-| 2 | | | | | | | |
-| 3 | | | | | | | |
-
-## Video Generation Prompts
-
-| Scene | Provider / CMS Node | Model | Prompt | Source Image | Duration | Camera Motion | Expected Cost |
-|---|---|---|---|---|---|---|---|
-| 1 | | | | | | | |
-| 2 | | | | | | | |
-
-## TTS Prompts
-
-| Scene | Provider | Voice | Text | Speed | Emotion | Expected Cost |
-|---|---|---|---|---|---|---|
-| 1 | | | | | | |
-| 2 | | | | | | |
-
-## Music Prompts
-
-| Track | Provider | Genre | Mood | Duration | Instrumental | Expected Cost |
-|---|---|---|---|---|---|---|
-| BG | | | | | | |
-
-## Style Consistency Rules
-
-<!-- Global style directives that apply across all prompts. -->
-
-- Base style: `<photorealistic / illustrated / anime / flat motion / minimalist>`
-- Color palette: `<description or hex values>`
-- Typography: `<font family, weight, treatment>`
-- Lighting: `<natural / studio / cinematic / neon>`
-- Consistency anchor: `<reference image URL or description>`