
mardab96/gtm-tracking-claude-skills
Developer toolsA pack of 15 production-ready Claude Skills for GTM, GA4, Google Ads, Meta Pixel+CAPI, consent, click IDs, duplicate events and CRM offline conversions.
Overview
A pack of 15 production-ready Claude Skills for GTM, GA4, Google Ads, Meta Pixel+CAPI, consent, click IDs, duplicate events and CRM offline conversions. Each skill is a self-contained SKILL.md with explicit triggers, required inputs, analysis workflow, decision rules, a practical example, output format, and guardrails. Drop them into Claude Code and they activate automatically when their use case matches the conversation. 1. Clone or download this repository. 2. Copy the skill folders into your project's .claude/skills/ directory, or into the user-level skills directory at ~/.claude/skills/. 3. Start a new Claude Code session in that project. Skills will activate automatically when their description matches the conversation. The skills are plain Markdown with YAML frontmatter and work anywhere Claude can read files. Paste the relevant SKILL.md content into context when you want to use it. Once installed, describe what you want in natural language.
README
15 Claude Skills for Google Tag Manager and Tracking
A pack of 15 production-ready Claude Skills for GTM, GA4, Google Ads, Meta Pixel+CAPI, consent, click IDs, duplicate events and CRM offline conversions.
Each skill is a self-contained SKILL.md with explicit triggers, required inputs, analysis workflow, decision rules, a practical example, output format, and guardrails. Drop them into Claude Code and they activate automatically when their use case matches the conversation.
What is inside
| # | Skill | Folder | Purpose |
|---|---|---|---|
| 1 | Gtm Container Health Check | gtm-container-health-check |
Reviews container structure, naming, versions and obvious errors. |
| 2 | Ga4 Event Quality Audit | ga4-event-quality-audit |
Checks event naming, parameters, key events and reporting logic. |
| 3 | Google Ads Conversion Tracking Audit | google-ads-conversion-tracking-audit |
Verifies conversion actions, values, attribution and primary/secondary setup. |
| 4 | Meta Pixel Capi Signal Quality Audit | meta-pixel-capi-signal-quality-audit |
Reviews Pixel, Conversions API, event quality scores and deduplication. |
| 5 | Consent Mode Impact Review | consent-mode-impact-review |
Evaluates consent setup and estimates signal loss impact. |
| 6 | Click Id And Utm Sanitization Check | click-id-and-utm-sanitization-check |
Checks gclid, fbclid, UTM handling across pages and redirects. |
| 7 | Duplicate Event Finder | duplicate-event-finder |
Finds events counted twice or fired from multiple sources. |
| 8 | Conversion Deduplication Review | conversion-deduplication-review |
Verifies deduplication logic for Google Ads and Meta. |
| 9 | Offline Conversion Import Audit | offline-conversion-import-audit |
Reviews CRM offline import setup and match quality. |
| 10 | Enhanced Conversions Readiness Check | enhanced-conversions-readiness-check |
Checks if Enhanced Conversions for Web or Leads are set up correctly. |
| 11 | Server Side Tagging Feasibility Review | server-side-tagging-feasibility-review |
Evaluates whether sGTM is worth it and what needs to move first. |
| 12 | Attribution Window Sanity Check | attribution-window-sanity-check |
Checks if attribution settings match your sales cycle and reporting needs. |
| 13 | Ecommerce Data Layer Audit | ecommerce-data-layer-audit |
Reviews purchase, item and checkout data layer quality. |
| 14 | Ios Privacy Signal Loss Review | ios-privacy-signal-loss-review |
Diagnoses impact of privacy changes on signal quality. |
| 15 | Weekly Tracking Signal Readout | weekly-tracking-signal-readout |
Produces a weekly tracking summary with confidence score and next actions. |
How to install
Option A — Claude Code (recommended)
- Clone or download this repository.
- Copy the skill folders into your project’s
.claude/skills/directory, or into the user-level skills directory at~/.claude/skills/. - Start a new Claude Code session in that project. Skills will activate automatically when their description matches the conversation.
git clone https://github.com/mardab96/gtm-tracking-claude-skills.git
mkdir -p ~/.claude/skills
cp -r gtm-tracking-claude-skills/*-* \
gtm-tracking-claude-skills/references \
~/.claude/skills/
Option B — Other Claude environments
The skills are plain Markdown with YAML frontmatter and work anywhere Claude can read files. Paste the relevant SKILL.md content into context when you want to use it.
How to use
Once installed, describe what you want in natural language. Skills self-trigger on the patterns documented in their ## Use this skill when section.
Activation and trigger reliability
Claude Skills usually activate from the YAML description field first, then use the full SKILL.md instructions after activation. Each skill in this pack now names the moment it should be used in both places:
descriptionexplains the tracking situation that should trigger the skill.## Use this skill whenlists concrete request patterns and decision moments.## Required inputtells the user what exports, screenshots, traces and reports make the skill reliable.
For best results, ask for the job and include the tracking object being diagnosed. Example: “Audit Google Ads conversion tracking before we scale” will trigger google-ads-conversion-tracking-audit more reliably than “Check tracking.”
Bring your data
Every skill is data-first. They work best when you provide actual exports, reports, screenshots or notes. Each SKILL.md documents exactly which inputs it expects under ## Required input.
Guardrails
These skills are designed for diagnosis and approval-ready recommendations. By default they do not:
- publish ads
- change budgets
- pause campaigns
- edit tracking setup
- change optimization events
- modify landing pages
- modify CRM fields
- send customer messages
- make performance claims without data
Use them to structure the work, not to hand over control of the account.
Validate the pack
Run the validator before publishing changes:
python3 scripts/validate-skills.py
It checks for placeholder text, missing required sections, weak activation descriptions and workflows that are too short to be useful.
Examples and review
Use the sample outputs to judge the expected depth of analysis:
Use evals/review-checklist.md for the next manual review pass.
License and attribution
Free to use, modify, and redistribute for personal and commercial work. If you ship something built on top of these skills, a credit is appreciated but not required.
Built and maintained by AdLume — performance marketing infrastructure for AI-native operators.
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add mardab96/gtm-tracking-claude-skills