aaron-marketing-skills is a collection of 120 AI-agent skills covering marketing work such as brand narrative, search optimization, social media, email, advertising, influencer campaigns, and launches. Marketers and agent users can install it as a plugin, use its portable skills, or run its described bot team. The catalogue entries are components of this marketing workflow.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/aaron-he-zhu/aaron-marketing-skillsnpx agentmods add skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qaWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qa)<a href="https://agentmods.dev/skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qa"><img src="https://agentmods.dev/badge/skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qa/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qa"><img src="https://agentmods.dev/badge/skills/aaron-he-zhu/aaron-marketing-skills/conversion-signal-qa.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk warn
- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 69 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00151 | $0.02311 |
| Opus 5 | $0.00076 | $0.01156 |
| Sonnet 5 | $0.00030 | $0.00462 |
| Haiku 4.5 | $0.00015 | $0.00231 |
Grade A, and why
conversion-signal-qa scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Conversion Signal QA
Pre-flight QA of the measurement plumbing behind paid ads — conversion-event firing, UTM hygiene, cross-platform dedup rules, attribution-window alignment, and offline/iOS-ATT modeled-gap flags — delivered as a tracking pre-flight checklist plus a UTM/event-spec builder. Scope line: this skill BUILDS and FIXES the signal pre-flight so the data is trustworthy; it does NOT score the ROAS R1/R2 vetoes — ad-account-auditor judges those as scored red lines. It is the R1/R2 prerequisite, not the verdict. It is also not the standing monthly de-dup / incrementality reconciliation — that is attribution-reconciler. Here you only gate that a dedup rule and aligned attribution windows exist pre-launch; the actual order-ID matching, double-count quantification, and incrementality read happen in attribution-reconciler.
Quick Start
QA my conversion tracking before I scale. Platforms: Google + Meta. Here is my GA4 Conversions export and Traffic-acquisition (source/medium) export: [paste/path].
Build me a UTM scheme and event spec for this campaign, then give me a pre-launch tracking checklist I can run myself.
My Meta and Google numbers don't match my GA4 orders — find the dedup, attribution-window, and UTM problems. [GA4 exports attached]
Skill Contract
Expected output: a tracking pre-flight checklist (pass/fail/needs-input per item), field-level evidence observations, a versioned UTM/event-spec binding (naming convention + conversion-event table + exact ref/hash), cross-platform dedup + attribution-window alignment notes, offline/iOS-ATT modeled-gap flags, and the standard handoff summary.
- Reads: site/account topic and platforms; the user's own GA4 Conversions report export and Traffic-acquisition (source/medium) export with source ref, observation time, window, currency, and timezone; one manual test conversion the user performs (NOT pixel/tag-manager API access).
- Writes: a user-facing pre-flight report plus a reusable UTM/event spec to
memory/ad/conversion-signal-qa/. - Promotes: signal-integrity blockers (events not firing, UTM gaps, dedup/window mismatch, missing test conversion) and the UTM/event spec to
memory/hot-cache.mdandmemory/open-loops.md. - Done when: every pre-flight item is marked pass/fail/needs-input from source- and time-bound evidence; the UTM scheme + event spec have a stable ref/version/hash; conflicting sources remain visible; dedup rules and attribution-window alignment are stated per platform; offline/iOS-ATT modeled gaps are flagged (never silently passed); and the report says the plumbing is launch-ready or names exactly what to fix.
- Primary next skill: ad-account-auditor to score
R1/R2and the full RQS once the signal is fixed.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago Changed · +3 lines 6964befb1838
- 13d ago First seen · 81 lines · 151 tokens per session scan A 6452e211ae2b
conversion-signal-qa is a skill published in the GitHub repository aaron-he-zhu/aaron-marketing-skills (2,767 stars, last pushed today), licensed Apache-2.0. It adds 151 tokens to every session and 2,311 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
geo-visibility-check
One-shot GEO audit: does your brand appear in Claude, ChatGPT, and Gemini answers for the buyer questions that matter? Runs a prompt panel through the engines with citation tracing and reports per-prompt verdicts, who wins instead, and which sources the answers come from.
geo-optimizer-skill
Run geo audit first. It scores the site 0–100 across 8 categories and generates a prioritized action list.
geo-loop
Run one bounded eGEOagents loop iteration over a workspace domain - read the charter and fresh collector data, do ONE unit of work, write substrate artifacts, append one Timeline entry and one LOG line. Use for loop mode, /geo:loop, scheduled GEO runs, or continuous monitoring.
content-scoring
Score content against the 10 GEO criteria with evidence and prioritized fixes. Use when users ask to score, rate, evaluate, or estimate ranking strength.
competitive-analysis
Analyze AI-search competitors for a query and recommend ranking strategy. Use when users ask competitor analysis, who ranks, or competitive landscape.
schema-generator
Generate JSON-LD schema markup for pages and content types with an implementation checklist. Use when users ask for schema, structured data, rich snippets, or markup.