Client context

Client Context — Pinterest Intelligence Layer

Purpose

Read the client's real documents and images to build the context layer that makes every composed pin client-specific. Pinterest is a visual search engine — context here means images and search terms as much as voice and audience. The image library inventory is the foundation of composed pin output.

Core Principle

This skill exists so that /create produces pins from real brand assets, not generic templates. Without context, pins are generic. Without images, pins are incomplete. Without learnings, pins don't improve. This skill loads all three.

When This Activates

Every /create and /track command starts here. Load context silently — show a one-line status header and move to the command's work. Don't narrate the loading process.

Priority Documents

Read the ~~docs folder. Look for these in priority order:

Priority	Document Type	What to Extract	If Missing
1	Image library	Available images — from ~~images (Cloudinary) or ~~docs images/ folder. Filenames, dimensions, orientation, categories, tags, visual descriptions. This is the most critical asset.	Cannot produce composed pins. Flag: "No images found. Add brand images to create composed pins."
2	Brand guide / style guide	Visual identity, voice for descriptions, vocabulary rules, banned words	Ask for brand name and 2-3 adjectives
3	Product / service catalog	Names, descriptions, prices, destination URLs, categories	Ask what they sell and where it links
4	Audience personas	Demographics, interests, Pinterest search behavior, planning patterns	Ask who the target Pinterest user is
5	Board structure	Board names, descriptions, themes, goals	Recommend boards based on brand vertical
6	Pin learning files	pin-log.md, pin-learnings.md, pin-baselines.md from /track	First batch — no learning history
7	Pinterest config	pinterest-config.md from /setup	Run /setup first or proceed with defaults
8	Content strategy	Goals, pillars, seasonal calendar, pin queue	Ask for primary goal

Image Library Scanning

This is the most critical context step for composed pin production.

Source Priority

1. ~~images (Cloudinary) — if connected, this is the primary image source. Advantages: automatic orientation data, metadata tags, transformation URLs for Pinterest sizing, CDN delivery.

2. ~~docs images/ folder — if Cloudinary is not connected, scan the Drive images folder. Less metadata available but functional.

3. Both — if both exist, Cloudinary is primary (better metadata and transforms), Drive is supplemental.

Image Inventory Protocol

For each image in the library, capture:

Dimension	What to Capture	Why It Matters
Filename / Asset ID	Exact reference for pin output	Required for composed pin attribution
Dimensions (px)	Width × height	Determines orientation and Pinterest suitability
Orientation	Vertical (2:3), square (1:1), horizontal (3:2)	Vertical gets 60% more engagement on Pinterest
Aspect ratio	Exact ratio	Determines if smart crop is needed for 2:3
Category	Product, lifestyle, team, diagram, quote card, seasonal, behind-scenes	Matches pin concept type
Tags / metadata	Cloudinary tags or Drive folder conventions	Enables topic-based search during /create
Visual description	What's in the image — subject, setting, mood	Enables concept-to-image matching
Quality signals	Resolution adequate? Well-lit? Clear subject?	Low-quality images hurt pin performance
Reuse status	Used in last 10 pins? (from pin-log.md)	Pinterest deprioritizes duplicate images
Transform URL	Cloudinary URL with Pinterest sizing applied	Ready-to-use for composed pin output

Orientation Classification

Orientation	Aspect Ratio	Pinterest Performance	Action
Vertical	2:3 or taller	Optimal — 60% more engagement	Use as-is or crop to exactly 1000×1500
Square	1:1	Acceptable — less feed space	Can use, or crop to vertical with smart crop
Horizontal	3:2 or wider	Worst performer	Smart crop to vertical if possible, flag if not

Image Health Assessment

After scanning, assess the library:

Condition	Status	Recommendation
20+ vertical images	🟢 Healthy	Sufficient for sustained batches
10-19 vertical images	🟡 Thin	Will run low within 2-4 weeks of regular posting
1-9 vertical images	🟠 Critical	Add more vertical images before regular production
0 images	🔴 Empty	Cannot produce composed pins. Blocker.
All horizontal	⚠️ Misfit	Pinterest penalizes horizontal. Add vertical shots.
No new images in 60+ days	⚠️ Stale	Pinterest rewards visual freshness. Add new photos.
Fewer images than batch size	⚠️ Insufficient	Reduce batch size or add images

Category Distribution

Track the mix of image types:

Category	Ideal For	Count	Status
Product shots	Product spotlight pins, sales goal
Lifestyle	Aspirational pins, save goal
Team / workspace	Behind-scenes, brand building
Diagrams / infographics	Educational, how-to pins
Quote cards / text overlay	Thought leadership
Seasonal	Timely, event-driven pins

If any category has zero images but the pin queue includes topics that need that category, flag: "Pin queue includes [topic] but no [category] images exist. Add [specific suggestion] before creating pins for that topic."

Brand Voice (Adapted for Pinterest)

Pinterest copy is shorter and more search-oriented than email or blog. Extract voice differently:

Dimension	Where to Find It	Pinterest Adaptation
Core visual identity	Brand guide	Informs image selection preferences and text overlay decisions
Description voice	Brand guide tone	Same voice, fewer words. Pinterest descriptions are 150-300 chars.
Vocabulary — use	Word list from guide	Prioritize search-relevant terms from the list
Vocabulary — avoid	Banned words from guide	Apply same restrictions in titles and descriptions
CTA style	Brand guide or past pins	Pinterest CTAs are soft — "Tap to explore" not "BUY NOW"

Critical Pinterest voice rule: Descriptions serve search first, personality second. Front-load keywords. Save the brand character for the title. A SaaS brand's description: "Content workflow automation for stretched marketing teams. Reduce approval cycles by 60% with editorial infrastructure." Not: "We believe content should be joyful!"

Audience Segments (Pinterest-Specific)

Pinterest users behave differently from other channels:

Dimension	What to Extract
Pinterest demographics	60% female (male growing fastest), median age 25-44, high-income skew
Search behavior	What terms does this audience type into Pinterest?
Planning stage	Planning ahead (saves) or ready to act (clicks)?
Save vs. click profile	Some audiences build boards. Some click through. Which is this?
Device	85% mobile. All pins must look good at mobile dimensions.

Intelligence Levels

Data Available	Intelligence Level
Personas only	Basic — demographics, interests, likely search terms
Personas + pin history	Moderate — knows which topics and image types engaged
Personas + history + learnings	Deep — knows save rate by image-copy combo, CTR by board
All above + analytics	Full — real data per variable

Board Structure

For each configured board:

• Name and description (descriptions should include keywords for Pinterest SEO)
• Theme — what content belongs here
• Goal — traffic, saves, awareness, sales
• Pin count — from pin-log.md
• Board health — getting fresh pins? Growing?
• Best image-copy combo — from learning files (if tracked)

Board strategy rules:

• Minimum 4 boards for serious presence
• Each board needs 15-25 pins to look established
• Board descriptions must include search keywords
• Seasonal boards can be archived post-season but pins stay discoverable

Performance Baselines

Source	What to Read
pin-baselines.md	Overall averages, per-board, per-image-type, per-combo baselines
pin-learnings.md	Cumulative learnings by category
pin-log.md	Last 20-30 pins with images used, boards, metrics

Confidence Scale

Tracked Pins	Confidence	What the System Knows
0	Baseline	Pinterest vertical benchmarks only
1-10	Low	Directional. "Lifestyle images seem to save better."
11-30	Medium	Board-level patterns. Image type preferences. Title effects.
31-75	High	Multi-variable patterns. Image-copy combo predictions.
76+	Very High	Full predictive capability. Specific combinations optimized.

Connector Awareness

Connector	If Available	If Unavailable
~~docs	Read all documents. Write learning files back. Full context.	Ask user to describe brand and paste info. No persistent learning.
~~images	Scan Cloudinary for image library. Get metadata, transforms, CDN URLs. Composed pins with ready-to-use images.	Fall back to ~~docs images/ folder. Less metadata, no transforms.
~~pinterest	Post pins via API. Pull analytics for /track.	User posts manually. Pastes analytics for /track.
~~calendar	Seasonal timing recommendations.	Use general seasonal knowledge.

Degradation Behavior

Level	Connected	Capability
Full	~~docs + ~~images + ~~pinterest	Complete: context + composed images + auto-post + auto-track
Composed	~~docs + ~~images	Strong: context + composed images. User posts manually.
Docs only	~~docs (with images/)	Good: context + image matching from Drive. No transforms.
Manual	Nothing	Functional but limited: user describes brand, provides images. No learning persistence.

Status Header

Display one line at the start of every command:

• 🟢 Config: Connected — [brand] — [N] images ([N] vertical) — [N] boards — [confidence level]
• 🟡 Config: Partial — [what's found] — missing [what's not] — [image library status]
• ⚪ Config: Manual — provide brand context and images

Cross-Plugin Compatibility

Same Drive folder serves Editorial OS, Email Marketing Strategist, and Pinterest Marketing Strategist. Shared: brand guide, audience personas, content strategy. Pinterest-specific: image library (shared with Composer), pin-log.md, pin-learnings.md, pin-baselines.md.

Teammate Protocol Compatibility

This skill works in both modes:

Plugin mode (Cowork): Reads from ~~docs and ~~images. User triggers /create and /track manually.

Teammate mode (Autonomous): Reads from workspace Drive folder. Cron triggers create and track per standing orders. Same skill, different trigger. Same learning files, same image library.

Standing orders example:
- Task: Create composed pin batch
  Day: Monday, Wednesday, Friday
  Skill: pin-strategist + client-context
  Image source: Cloudinary folder "editorial-os/pinterest"
- Task: Track pin performance
  Day: Monday
  Skill: performance-learning