amekala/ads-mcp/adspirer-setup
Claude Code and agent skill by amekala/ads-mcp
- community
Checking that the SKILL.md file is still in the repository.
Set up your brand workspace — connect to Adspirer, scan brand docs, pull campaign data, and create CLAUDE.md. Use on first time in a new brand folder, or when the user wants to refresh brand context.
Install
npx vskill@latest install amekala/ads-mcp/adspirer-setupSKILL.md
Run the full brand workspace setup. Follow these steps in order:
Step 1: Connect to Adspirer
Call get_connections_status directly.
- If it works: continue to step 2.
- If the MCP server is not found: tell the user:
"The Adspirer MCP server isn't connected. Add it to your Cursor MCP config:
- Open Cursor Settings > MCP
- Add server with URL:
https://mcp.adspirer.com/mcp - Or add to
~/.cursor/mcp.json:
``json {"mcpServers": {"adspirer": {"url": "https://mcp.adspirer.com/mcp"}}} ``
- Restart Cursor and run
/adspirer-setupagain."
- If OAuth is triggered: tell the user a browser window is opening for authentication. Wait for them to confirm, then retry.
- If no platforms are connected: tell the user to connect ad accounts at https://adspirer.ai/connections, then run setup again.
Step 2: Scan local folder
Search for brand documentation files: **/*.md, **/*.txt, **/*.csv, **/*.yaml, **/*.json, **/*.pdf
Read all discovered files. Extract:
- Brand name and industry
- Products/services
- Target audience descriptions
- Brand voice and tone indicators
- Competitor mentions
- Budget information
- KPI targets or performance goals
- Previous campaign strategies or results
If the folder is empty, that's fine — we'll build context from Adspirer data alone.
Step 3: Pull live campaign data
Follow adspirer-mcp for the call contract. Only these are callable by name; everything else goes through its platform router with {"action": "list_tools"} then {"action": "execute", ...}.
Call directly:
get_connections_status— which platforms are connected. Do this first; only pull the rest for
platforms that came back connected.
get_campaign_performance(lookback_days: 30) — Google Ads performanceget_meta_campaign_performance(lookback_days: 30) — Meta performance
Through the router (action: "execute"):
google_ads→list_campaigns,get_business_profile,get_benchmark_contextlinkedin_ads→get_linkedin_campaign_performance(lookback_days: 30)
If any tool errors (platform not connected), skip it and note the gap.
Microsoft Advertising, when microsoft_ads is connected, has list tools for campaigns and budgets and no performance report unless list_tools shows one. Record status and budgets. Metrics are "not available", not 0. See adspirer-microsoft-ads.
If a performance read returns no data, record freshness_state and the range that actually has data. Don't write 0 for a metric you didn't receive, and don't add a blended conversions or blended CPA row across platforms or currencies.
Business type, before you invent a goal
If the docs and the account don't already say it, ask once what the business is: lead gen or B2B sign-up, ecommerce, app installs, calls or local, or awareness. That answer is the primary goal. Later tracking checks judge conversion actions against it.
Flag accounts whose names look like test, sandbox, demo, or an end-to-end fixture, and ask which are real before you treat their spend as the brand.
Step 4: Create CLAUDE.md
Existing-file guard — check before writing:
- No
CLAUDE.mdin the project root: create it (the user explicitly asked for setup, which covers this). CLAUDE.mdexists and is already an Adspirer brand workspace: update it in place, preserving any edits the user made.CLAUDE.mdexists and is anything else (e.g. a software project's instructions): NEVER overwrite, restructure, or delete it. Ask the user whether to append a clearly-marked## Adspirer Brand Contextsection at the end, or to skip the file and keep brand context in the conversation only. Do not touch the file until they answer.
Generate the brand workspace by combining local files + Adspirer data:
Fill in every section with real data. Leave placeholders only for sections where no data was found — and note the gap.
Step 4.5: Create STRATEGY.md
Create STRATEGY.md at the project root alongside CLAUDE.md:
Step 5: Create memory file
Create .cursor/memory/performance-marketing-agent/MEMORY.md with initial template:
Step 6: Present summary
Tell the user:
- Which platforms are connected and how many campaigns are active
- A quick performance snapshot (spend, CTR, CPA, ROAS)
- Key findings (top campaigns, wasted spend, opportunities)
- Any gaps ("No brand voice docs found — drop guidelines in this folder anytime")
End with: "Your brand workspace is set up! I've saved everything to CLAUDE.md. Here's what I can help with:
- Review campaign performance across all platforms
- Check conversion tracking in the ads manager (read-only) before trusting a CPA
- Find and fix wasted ad spend
- Write brand-voice ad copy
- Create new campaigns (they stay paused until you say to enable them)
- Research keywords
- Set up monitoring and alerts
What would you like to start with?"
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/adspirer-setup/SKILL.md
Popularity
- Repository stars
- 79
- Repository forks
- 12
- 7-day trend score
- 22
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 6 Jun 2026, 5 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/amekala/ads-mcp/adspirer-setup)