Durga-veer

Elevate Your Campaign — AI skill for Claude Code

AI community

AI marketing campaign diagnosis agent.

How to install Elevate Your Campaign

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Durga-veer/Elevate-Your-Campaign and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Elevate Your Campaign does

AI marketing campaign diagnosis agent. Upload a campaign CSV and it finds anomalies, explains the root cause with evidence and confidence, recommends next steps, designs an experiment, and remembers what worked. Deterministic TypeScript engine does all the math; Claude only explains and a Critic agent fact-checks.

Alternatives in AI

Cursor Talk To Figma MCP — TalkToFigma: MCP integration between AI Agent (Cursor, Claude Code, Codex) and Figma, allowing Agentic AI to c 7k ★
Osint Skill — OSINT Skill for AI agents (Claude Code, OpenClaw, Codex, OpenCode) — from a name to a scored dossier with psyc 114 ★
Aurora — Aurora - independent community smart home orchestrator 103 ★

README

⚡ Elevate Your Campaign: AI Marketing Campaign Diagnosis Agent

Most tools tell you what happened. This one tells you **why**, **what to do next**, **tests it**, and **remembers what worked**.

**DATA → ANOMALY → ROOT CAUSE → EVIDENCE → DIAGNOSIS (confidence) → RECOMMENDATION → ACTION PLAN → EXPERIMENT → RESULT → MEMORY**

Run it in VS Code

npm install
cp .env.example .env      # add ANTHROPIC_API_KEY (optional: without it, offline templates are used)
npm run dev               # http://localhost:5173

Click **⚡ ELEVATE MY CAMPAIGN**. With no data uploaded, it uses the built-in sample with 3 planted problems and 1 opportunity.

Other scripts:

npm run build: typecheck and production build
npx vite-node scripts/smoke-test.ts: checks the engine finds every planted problem and the Critic passes
node scripts/api-test.mjs: end-to-end API test against a running dev server
npm run gen:csv: writes public/sample_campaign_data.csv

Any CSV: two analysis modes (picked automatically by the Data Agent)

Your file	Mode	What you get
Daily export (`date, impressions, clicks, spend`, plus campaign, conversions, revenue…)	Daily trends (`engine/diagnosis.ts`)	Change points, metric-tree root causes (fatigue, landing page, tracking, auction), budget waste, audience opportunities
No dates (e.g. one row per customer or per campaign) with a category column and a spend/conversion column	Segment (`engine/segments.ts`)	Segment comparison with significance tests (z-test, p < 0.01), weak/strong segments, conversion drivers (converted vs not, Cohen's d), age/income bands

Check any file offline: `npx vite-node scripts/analyze-file.ts path/to/file.csv`

Architecture

Two layers. **The LLM never does the math.**

Layer	Where	What
Deterministic engine (TypeScript)	`src/shared/engine/`	CSV parsing, metrics, period comparison, z-score change points, metric-tree rules, priority, health, number