creative-angle-lab

name	creative-angle-lab
description	When the user wants new paid social creative angles, hook banks, UGC briefs, TikTok or Meta ad concepts, creative testing matrices, performance creative strategy, offer angles, customer objection angles, top-spending creative analysis, Gemini image or video breakdowns, awareness-level mapping, avatar mapping, or ways to turn reviews and ad data into new tests. For finished scripts, use ugc-script-writer.
metadata	{"version":"1.0.0"}

Creative Angle Lab

Turn context, performance data, reviews, objections, and competitor observations into testable creative angles.

Before Ideation

1. Read .agents/vibemarketer-context.md if it exists.
2. Pull in performance findings from meta-performance-analyst if available.
3. Ask for or inspect customer reviews, comments, support tickets, top-spending ads, losing ads, landing page copy, and competitor examples when useful.
4. Separate proven facts from hypotheses.

Reference

Load only what is needed:

• references/angle-frameworks.md for angle types and testing structures.
• references/gemini-creative-analysis.md for full image/video creative breakdowns using Gemini.

Scripts

Use these when the user wants executable LLM creative analysis:

• scripts/build_creative_handoff.py: convert normalized Meta ad rows into a top-creative handoff for this skill, with optional Meta Ads CLI creative metadata and asset downloads.
• scripts/analyze_creative_gemini.py: send image or video assets to Gemini.
• scripts/analyze_creative_batch_gemini.py: analyze a top-creative handoff CSV with Gemini and synthesize the portfolio angle map.
• scripts/analyze_static_creative_openai.py: analyze static images with OpenAI Responses API.
• scripts/analyze_static_creative_claude.py: analyze static images with Claude Messages API.

Gemini remains the default for video. OpenAI and Claude are static-creative fallbacks unless the user asks otherwise.

Workflow

1. Extract raw material

   • Customer pains
   • Desired outcomes
   • Objections
   • Proof points
   • Product mechanisms
   • Competitor gaps
   • Performance signals
   • Top-spend creative assets and transcripts

2. Build angle territories

   • Group ideas by customer motivation, not by minor wording.
   • Avoid duplicates disguised as new angles.

3. Prioritize

   • Score each angle on relevance, proof strength, novelty, production ease, and risk.
   • Favor angles that can be tested with multiple hooks and creator types.

4. Output test plan

   • Angles
   • Hooks
   • Creator brief
   • Visual concept
   • Proof asset
   • CTA
   • Success metric

Top-Spend Creative Analysis

When creative assets are available, do not rely only on thumbnail/frame sampling. Send images or videos to Gemini for a full visual breakdown when the environment has Gemini access and user consent.

Use meta-performance-analyst as the upstream triage layer. It should identify top-spend creatives, high-CPMr watch items, efficient reach candidates, and campaign-level dependencies. This skill should then explain the creative patterns behind those signals and produce the next angle map. Keep media-buying allocation decisions in meta-performance-analyst; keep hook, avatar, awareness-level, and creative expansion decisions here.

When starting from Meta output, build the handoff first:

python3 skills/creative-angle-lab/scripts/build_creative_handoff.py \
  --ads data/live-analysis-pack-v3-conversions/meta-active-hierarchy/ads_normalized.csv \
  --output-dir outputs/creative-angle-lab/demo-brand \
  --resolve-meta-creatives \
  --download-assets

Then run the Gemini portfolio read:

python3 skills/creative-angle-lab/scripts/analyze_creative_batch_gemini.py \
  --handoff-csv outputs/creative-angle-lab/demo-brand/creative-handoff.csv \
  --output-dir outputs/creative-angle-lab/demo-brand/gemini-analysis

Default artifact structure:

• portfolio-synthesis.md: standalone executive read for fast sharing.
• creative-angle-lab-report.md: per-creative report only; it should not duplicate the full portfolio synthesis.
• per-creative/*.md: individual Gemini reads for each analyzed ad asset.

Use --download-assets only when the account token can access downloadable image or video URLs. If only thumbnails or metadata are available, label visual analysis as incomplete and use the handoff to request/source the original assets. For videos, prefer the downloaded MP4. For static images, prefer the ad-image hash URL over thumbnail_url.

In synthesis outputs, cite actual Meta ad IDs and short names. Avoid ordinal-only references like "Creative 2" because they are hard to trace back to Ads Manager.

For each top-spending creative, classify:

• Angle
• Hook
• First 3 seconds
• Awareness level
• Avatar
• Core desire
• Core pain
• Mechanism
• Proof type
• Offer frame
• Visual style
• Creator archetype
• CTA
• Comments or objections if supplied

Then produce a portfolio read:

• What the account is over-indexed on.
• What awareness levels are missing.
• Which avatars are under-served.
• Which winning motifs should be expanded.
• Which themes are fatigued or too repetitive.

Default Output

## Angle Map
| Priority | Angle | Customer Insight | Proof Needed | Risk | Why Test |
| --- | --- | --- | --- | --- | --- |

## Hook Bank
| Angle | Hook | Format | Notes |
| --- | --- | --- | --- |

## Creative Test Matrix
| Test | Angle | Creator | Format | Hook | Proof | CTA | KPI |
| --- | --- | --- | --- | --- | --- | --- | --- |

## Creative Portfolio Diagnosis
| Pattern | Evidence | Coverage | Gap | Expansion |
| --- | --- | --- | --- | --- |

## Production Briefs
### Brief 1: [Angle]
- Creator:
- Opening visual:
- Product moment:
- Proof:
- Watchouts:
- Variants:

Quality Bar

• Make angles specific enough to shoot.
• Include the customer belief each angle is trying to change.
• Include proof requirements.
• Avoid making claims the brand cannot support.
• Do not generate 50 shallow ideas when 8 strong testable angles are better.