# Audit paid advertising accounts before scaling

Paid advertising decisions can rely on inflated attribution, incomplete measurement, or unsafe placements. This skill reviews exported evidence against a structured ROAS framework and returns a clear gate decision.

## Install

```bash
npx skillstore add aaron-he-zhu/ad-account-auditor
```

## Metadata

- Status: approved
- Slug: aaron-he-zhu-ad-account-auditor
- Version: 19.0.0
- Author version: 19.0.0
- Skillstore revision: r4
- Version status: valid
- Tree hash: 8710236724f0a73cb218ca972a3f62325af231dac88afc2a941a56e0667cd5c9
- Author: aaron-he-zhu
- GitHub username: aaron-he-zhu
- License: Apache-2.0
- Repository: https://github.com/aaron-he-zhu/aaron-marketing-skills/tree/main/ad/activate/ad-account-auditor
- Ref: adcb3549b15782055d0beb6d29f113d20de08f92
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, filesystem
- Quality score: 77
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/aaron-he-zhu-ad-account-auditor
- Manifest: https://skillstore.pages.dev/api/skills/aaron-he-zhu-ad-account-auditor/manifest

## Capabilities

- Collects required account context, including currency, attribution window, conversion lag, and business constraints.
- Scores 20 ROAS criteria across audience, offer, return, and spend efficiency.
- Separates platform-reported results from normalized own-data outcome evidence.
- Flags verified measurement, attribution, claims, policy, and placement-safety vetoes.
- Returns SHIP, FIX, BLOCK, or UNDECIDED outcomes based on evidence coverage and scoring.
- Produces prioritized fixes with owners and rerun conditions without changing account settings.

## Use Cases

- Approve budget growth: Review measurement and incrementality evidence before increasing spend on a paid-media portfolio.
- Investigate weak ROAS: Identify likely attribution, duplication, placement, or economics issues in an underperforming account.
- Prepare an audit handoff: Create an evidence-based account review with gaps, priorities, owners, and rerun conditions.

## Prompt Templates

### Start a basic audit

```
Audit this paid advertising account for direct response. I will provide campaign exports, conversion data, currency, attribution window, conversion lag, and CAC limit.
```

### Check launch readiness

```
Review this account before launch. Use the direct-response profile and identify any missing evidence, verified vetoes, and the correct SHIP, FIX, BLOCK, or UNDECIDED outcome.
```

### Assess incrementality

```
Audit this portfolio with the incremental-profit profile. Compare the holdout results and deduplicated order-ID exports against the stated margin, payback, attribution window, and conversion lag.
```

### Request an executive report

```
Run a complete ROAS audit using the supplied exports. Present business impact, evidence quality, 20-item scoring, verified controls, unknowns, prioritized owners, and a rerun condition. Do not change account settings.
```

## Limitations

- It evaluates evidence supplied by the user and cannot independently verify missing exports or source systems.
- It does not enable campaigns, edit bids, upload audiences, or change advertising budgets.
- It does not replace legal, platform-policy, finance, or experimental-design review.
- Standalone installs may collect observations but cannot calculate a score or persist an audit artifact.

## Best Practices

- Provide normalized exports with dates, IDs, currency, attribution windows, and conversion lag.
- Use own-data outcomes and holdout evidence when available instead of relying only on platform reporting.
- Treat missing required evidence as Unknown and resolve it before making scaling decisions.

## Anti Patterns

- Do not treat platform-reported ROAS as proof of incremental profit.
- Do not compare channels before normalizing currency, windows, lag, and outcome IDs.
- Do not change campaigns or budgets based on an audit without separate explicit approval.

## Security Audit

- Audited at: 2026-07-26T09:51:39.879\+00:00
- Summary: All 39 static findings are false positives caused by Markdown code formatting, static documentation links, metadata URLs, and fixed reference paths. The reviewed instructions limit the skill to assessment, require explicit authorization before persistence, and prohibit ad-account mutations without separate approval. No prompt injection, credential collection, data exfiltration, or arbitrary command execution evidence was found.

## Stats

- Views: 0
- Downloads: 5
- Favorites: 0
- Popularity score: 0