Ad Creative Edit logo

Ad Creative Edit

Organization
appeeky
ad-creative-edit

When the user wants to edit an existing Appeeky-generated ad creative with natural language instructions. Use when the user mentions "edit ad creative", "change background", "bigger headline", "iterate on ad image", "tweak the ad", or mode edit for app ad creatives. For fresh angles see meta-ad-creative. For publishing see ad-creative-to-meta.

Overview

Publisherappeeky
Repositoryua-skills
Skill namead-creative-edit
Stars
73
Forks
5
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by appeeky on GitHub. Read the source before you install it.

Installation

Install the Ad Creative Edit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/appeeky/ua-skills/tree/main/skills/ad-creative-edit
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/appeeky/ua-skills.git /tmp/ua-skills
mkdir -p .claude/skills
cp -r /tmp/ua-skills/skills/ad-creative-edit .claude/skills/ad-creative-edit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ad Creative Edit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Ad Creative Edit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Ad Creative Edit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Ad Creative Edit (Iterate)

Iterate on a previous Appeeky-generated ad using mode: edit. Use edits for incremental improvements on a winning or near-winning concept. Use fresh mode: generate for new angles, presets, or formats.

Editing is cheaper than guessing. Winners get A′/B′ variations; losers get retired, not endlessly patched.

When to Edit vs Re-Generate

SituationUse editUse generate
Bigger headline / font size✅
Background color / contrast✅
Minor layout tweak✅
Localize on-image text✅
New hook / new angle✅
Different preset (UGC → split)✅
Structural format change✅
Failed concept (low CTR + low CPA)✅ new variant

Prerequisites

InputSource
previous_ad objectget_app_ad_creative_job → result.ad
Same app_urlOriginal generation
Specific edit_instructionUser request — must be concrete
Optional key_pointsPreserved from analyze step

previous_ad Fields (API result object, camelCase)

json
{
  "style": "ugc",
  "imagePreset": "person_holding_phone",
  "creativeHeadline": "Plan your week in 5 min",
  "creativeSubhead": "The focus app that sticks",
  "imageUrl": "https://..."
}

If user has no job ID, reconstruct from Ads Manager export or screenshot description — quality drops without imageUrl.

Edit Instruction Framework

Good instructions are specific, visual, and bounded.

Instruction Template

Change [element] to [target state].
Keep [what must not change].

Examples

QualityInstruction
✅ Good"Make the on-image headline 30% larger, bold white text on navy background. Keep phone mockup position unchanged."
✅ Good"Switch background from gradient to solid charcoal. Increase contrast on subhead."
✅ Good"Translate on-image text to German. Keep layout identical."
❌ Bad"Make it better"
❌ Bad"More professional" (too vague)
❌ Bad"Copy competitor X" (use teardown + generate)

Workflow

Step 1 — Load Previous Creative

get_app_ad_creative_job
  job_id: "<original-job-id>"

Extract result.ad for previous_ad. Show user current image + copy before editing.

Step 2 — Clarify Instruction

If user request is vague, offer 2–3 concrete options:

Option A: Dark background + larger headline
Option B: Minimal white background, smaller subhead
Option C: Re-generate with professional preset (new job)

Step 3 — Call Edit API

MCP:

generate_app_ad_creative
  app_url: "https://apps.apple.com/us/app/id123456789"
  mode: "edit"
  edit_instruction: "Make headline bigger, navy background, more contrast on CTA text"
  previous_ad: {
    "style": "ugc",
    "imagePreset": "person_holding_phone",
    "creativeHeadline": "Plan your week in 5 min",
    "creativeSubhead": "The focus app that sticks",
    "imageUrl": "https://..."
  }
  quality: "medium"

REST:

bash
POST /v1/app-ad-creatives/generate

{
  "appUrl": "https://apps.apple.com/us/app/id123456789",
  "mode": "edit",
  "editInstruction": "Make headline bigger, navy background, more contrast",
  "previousAd": { ... },
  "quality": "medium"
}

Step 4 — Poll Job

get_app_ad_creative_job
  job_id: "<new-job-id>"

Poll every 3–5 seconds until completed or failed.

Step 5 — Present Before / After

Use output template. Ask user to pick version for Meta upload.

Edit Patterns That Work

RequestTypical resultNotes
Bigger headlineLarger type, same layoutSpecify % or "dominant"
Navy / dark backgroundHigher contrastGood for feed thumb-stopping
More minimalReduced clutterMay remove decorative elements
Brighter / warmerLifestyle feelPairs with lifestyle style
Localize textTranslated on-image copySpecify language code
Show different screenshotPartial changeMay need MCP extra_screenshot_urls (REST extraScreenshotUrls) + generate
Add rating starsSocial proof overlayOnly if accurate per listing

Edit Patterns That Fail

RequestWhyAlternative
"Make it viral"Not a visual parameterNew angle via generate
Change aspect ratioAPI outputs 1024×1024Crop in design tool
Add video motionStatic APItiktok-creative-strategy
Copy exact competitor layoutIP/policy riskTeardown-inspired generate
Fix fundamentally weak hookCopy problemNew angle in generate

Winner Iteration Playbook (A′ / B′)

After meta-campaign-audit identifies a winner:

Iterationedit_instruction focusGoal
A′ hookChange on-image headline text onlyIsolate hook
A′ visualBackground / contrast tweakCTR lift
A′ proofAdd rating or user countCVR lift
A′ localizeTranslate for EU geo testGeo expansion

Rule: One change per edit job. Multi-variable edits obscure what worked.

# Winner A′ — hook tweak only
generate_app_ad_creative
  app_url: "https://apps.apple.com/us/app/id123456789"
  mode: "edit"
  edit_instruction: "Change headline to 'Stop juggling 4 apps.' Keep all visual styling identical."
  previous_ad: { ...winner A... }

Copy-Only Refresh

If image is fine but Meta copy underperforms:

generate_app_ad_creative
  app_url: "https://apps.apple.com/us/app/id123456789"
  mode: "analyze"
  generate_image: false
  angle: "new hook: outcome-first for busy parents"

Then manually update Ads Manager text — or re-generate image if on-image text must match.

Credit Costs

QualityCreative credits
low1
medium2
high5

Same as generate. Inform user before each edit round.

Typical iteration session: 2–3 edits × medium = 4–6 creative credits.

Output Template

markdown
# Ad Creative Edit — [App Name]

## Edit Request
"[user instruction]"

## Before
- **Image:** [previous imageUrl]
- **Headline:** [previous creativeHeadline]
- **Subhead:** [previous creativeSubhead]
- **Primary text:** [previous copy.primaryText]

## After
- **Image:** [new imageUrl]
- **Headline:** [new creativeHeadline]
- **Subhead:** [new creativeSubhead]
- **Primary text:** [new copy.primaryText]

## Copy Changes (if any)
| Field | Before | After |
|-------|--------|-------|
| Primary text | | |
| Headline | | |

## Recommendation
- [ ] Ship to Meta (replace ad or new ad in same set)
- [ ] Edit again (specify next change)
- [ ] Re-generate fresh angle

## Meta Action
- Replace creative: upload new image → new ad OR swap in ad-creative-to-meta pipeline
- Keep loser paused — do not delete (learning history)

Meta Upload After Edit

Edited image must be re-uploaded to Meta:

meta_ads_upload_ad_image_from_url
  ad_account_id: "act_xxx"
  image_url: "<new imageUrl from edit result>"

Then create new ad with new imageHash — see ad-creative-to-meta. Pausing old ad preserves audit trail.

Benchmarks (edit impact)

MetricExpectation
CTR lift from contrast edit+10–25% if original was low contrast
CTR lift from hook-only edit+15–40% if new hook stronger
CPA changeFollows CTR × CVR — measure 48h
Diminishing returnsAfter 3 edits, re-generate new angle

Common Mistakes

MistakeFix
Editing losers instead of killingPause at 2× CPA; edit winners only
Stacking 5 edits on one adOne variable per edit
No imageUrl in previous_adAlways save job results
Editing without audit dataLet data pick the winner first
Same primary text, new image onlyAlign on-image and feed copy

Pre-Edit Checklist

- [ ] Original job_id and previous_ad saved
- [ ] User instruction is specific and bounded
- [ ] Credit cost communicated
- [ ] Confirmed edit vs re-generate decision
- [ ] Poll completed before presenting
- [ ] Meta re-upload planned if going live

Related Skills

  • meta-ad-creative — fresh generation from listing
  • ad-creative-variants — initial batch before edits
  • ad-creative-to-meta — publish edited creative
  • meta-campaign-audit — identify which ad deserves A′
  • competitor-ad-teardown — inspiration for new angles (not edits)
  • screenshot-optimization (aso-skills) — if edit needs new screenshot source

Frequently asked questions

What does the Ad Creative Edit AI skill do?

When the user wants to edit an existing Appeeky-generated ad creative with natural language instructions. Use when the user mentions "edit ad creative", "change background", "bigger headline", "iterate on ad image", "tweak the ad", or mode edit for app ad creatives. For fresh angles see meta-ad-creative. For publishing see ad-creative-to-meta.

Why use Ad Creative Edit on TypingMind?

Because you install it once and use it with any model. Ad Creative Edit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Ad Creative Edit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/appeeky/ua-skills/tree/main/skills/ad-creative-edit. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Ad Creative Edit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Ad Creative Edit?

As many as you like. As long as a model supports skills, you can use Ad Creative Edit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Ad Creative Edit AI skill free?

It is published on GitHub by appeeky. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇