Conversion Signal QA
Pre-launch QA of paid-ad tracking: UTM hygiene, conversion events, cross-platform dedup.
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 20, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 264f248
Fetched the tree via GitHub API and read SKILL.md plus both references (preflight-checklist.md, utm-event-spec.md) raw; all body-referenced files exist (both local references and the cross-skill links to ad-account-auditor, attribution-reconciler, skill-contract.md, roas-benchmark.md, CONNECTORS.md, SECURITY.md). Frontmatter parses with name+description; no hardcoded paths, no curl|sh/binaries/base64/exfiltration, and it treats imported CSVs as untrusted data. Baseline gave generic "check UTMs and watch double-counting" advice; the skill produced a structured pass/fail/needs-input pre-flight, caught the Google/google casing collision specifically, named GA4 order IDs as single source of truth, emitted filled UTM + event-spec tables with dedup IDs, and correctly framed iOS-ATT modeled data as a flag-not-fail — a clear, specific win, hence output 8/10. Trigger phrasings tested (all 5 correct): SHOULD — "QA my conversion tracking before I launch Google + Meta", "My Meta and Google numbers double-count, set the dedup rule", "Build me a UTM scheme + event spec and a pre-launch checklist"; SHOULD NOT — "Score my account's ROAS, does it pass R1/R2" (routes to ad-account-auditor), "Design my campaign/account structure" (campaign-architect). The description draws these boundaries explicitly and honestly.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 8/10
- Docs & honesty 5/5
What Conversion Signal QA does
A pure-prompt skill that runs a pre-flight QA of the measurement plumbing behind paid ads — conversion-event firing, UTM naming hygiene, cross-platform dedup rules, attribution-window alignment, and iOS-ATT/offline modeled-gap flags — from the user's own GA4 exports and one manual test conversion. Triggers when a user asks to QA their conversion tracking before launch/scale, fix UTMs, or set a dedup rule so Meta and Google stop double-counting. It gates whether the plumbing is launch-ready and hands the actual scoring/reconciliation to sibling skills.
How to install Conversion Signal QA
Plugin skill in a Claude Code marketplace repo (.claude-plugin/marketplace.json present). Copy ad/activate/conversion-signal-qa/ (SKILL.md + references/) into your skills directory, or add the marketplace and enable it. Self-contained: pure prompt + two reference markdown files, no external CLI/MCP required — ad-platform APIs are explicitly optional.
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Conversion Signal QA
-
/conversion-signal-qaPre-launch QA of paid-ad tracking: UTM hygiene, conversion events, cross-platform dedup.
It also activates on plain-language prompts like these:
-
QA my conversion tracking before launch -
Check my UTMs and pixel firing setup -
Stop Meta and Google from double-counting conversions
Frequently asked questions
- Is the Conversion Signal QA skill free?
- Yes. The skill itself is free from aaron-he-zhu/aaron-marketing-skills. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Conversion Signal QA work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 20, 2026. Verdict: Tested · Works. Fetched the tree via GitHub API and read SKILL.md plus both references (preflight-checklist.md, utm-event-spec.md) raw; all body-referenced files exist (both local references and the cross-skill links to ad-account-auditor, attribution-reconciler, skill-contract.md, roas-benchmark.md, CONNECTORS.md, SECURITY.md). Frontmatter parses with name+description; no hardcoded paths, no curl|sh/binaries/base64/exfiltration, and it treats imported CSVs as untrusted data. Baseline gave generic "check UTMs and watch double-counting" advice; the skill produced a structured pass/fail/needs-input pre-flight, caught the Google/google casing collision specifically, named GA4 order IDs as single source of truth, emitted filled UTM + event-spec tables with dedup IDs, and correctly framed iOS-ATT modeled data as a flag-not-fail — a clear, specific win, hence output 8/10. Trigger phrasings tested (all 5 correct): SHOULD — "QA my conversion tracking before I launch Google + Meta", "My Meta and Google numbers double-count, set the dedup rule", "Build me a UTM scheme + event spec and a pre-launch checklist"; SHOULD NOT — "Score my account's ROAS, does it pass R1/R2" (routes to ad-account-auditor), "Design my campaign/account structure" (campaign-architect). The description draws these boundaries explicitly and honestly.
- What is the Conversion Signal QA SkillProof Score?
- 9.2/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 8/10, docs & honesty 5/5.
- How do I install Conversion Signal QA?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Conversion Signal QA with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.