Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add metrikia-io/marketing-skills --skill conversion-verifiergit clone --depth 1 https://github.com/metrikia-io/marketing-skillsWrote 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/metrikia-io/marketing-skills/conversion-verifier)<a href="https://agentmods.dev/skills/metrikia-io/marketing-skills/conversion-verifier"><img src="https://agentmods.dev/badge/skills/metrikia-io/marketing-skills/conversion-verifier/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/metrikia-io/marketing-skills/conversion-verifier"><img src="https://agentmods.dev/badge/skills/metrikia-io/marketing-skills/conversion-verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00160 | $0.05108 |
| Opus 5 | $0.00080 | $0.02554 |
| Sonnet 5 | $0.00032 | $0.01022 |
| Haiku 4.5 | $0.00016 | $0.00511 |
Grade A, and why
conversion-verifier 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 11d 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 — 429 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Conversion Verifier
Ad platforms grade their own homework. Meta decides which sales Meta caused, Google decides which sales Google caused, and neither is ever asked to check its answer against the bank. This does that check.
The idea that makes this useful
Anyone can subtract two numbers and announce that the platform is lying. That tool would be wrong most of the time, because most of the gap between claimed and actual has legitimate causes: attribution windows shift dates, timezones slide orders across midnight, refunds are never reported back, and a store's total includes channels the platform never touched.
An experienced buyer knows all of this and will dismiss a naive "you're being lied to" report in about ten seconds.
So the work here is not the subtraction. It is separating the explained gap from the unexplained one. The unexplained residual is the only number worth acting on, and it is the number nobody currently has.
Hold that standard throughout. Being right matters more than being alarming, and a report that survives scrutiny is worth more to the person reading it than one that impresses them for a minute.
Checklist
Create a task for each of these and work through them in order.
- Get the ad data with the least friction that works: the Meta connection if they have it, the export if not
- Ask for the store export and the four questions, gross margin first
- Run
reconcile.pywith--gross-margin, deterministic numbers, no interpretation - Read the JSON as an analyst, using
references/gap-taxonomy.md - Deliver the finding in the chat first, then write the narrative to the caps
- Build the visual report with
build_report.py, then tell them how to get a PDF - Close with the limit and the offer, in that order, never reversed
Two rules that decide whether this gets used at all
Gaetan, a media buyer, tested the first version and gave two verdicts worth holding onto. Both are about friction rather than analysis.
What ships with it
18 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.
- examples/meta_export_totals.csv 389 B
- examples/meta_export.csv 5.3 KB
- examples/sample-report-FR.html 16 KB
- examples/sample-report.html 15 KB
- examples/shopify_orders.csv 105 KB
- README.md 7.6 KB
- references/gap-taxonomy.md 5.6 KB
- references/how-to-export.md 4.6 KB
- references/report-template.md 4.8 KB
- scripts/build_report.py 6.1 KB runs code
- scripts/charts.py 13 KB runs code
- scripts/columns.py 5.5 KB runs code
- scripts/gap_analysis.py 7.0 KB runs code
- scripts/loaders.py 7.0 KB runs code
- scripts/reconcile.py 26 KB runs code
- scripts/report_html.py 32 KB runs code
- scripts/safe_html.py 3.9 KB runs code
- tests/test_pipeline.py 17 KB runs code
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.
- 11d ago First seen · 429 lines · 160 tokens per session scan A 04d8a6fe4b7d
conversion-verifier is a skill published in the GitHub repository metrikia-io/marketing-skills (1 stars, last pushed 1mo ago), licensed MIT. It adds 160 tokens to every session and 5,108 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 2026-08-31.
Other skills, from other repositories
Threat Intelligence & CTI
Cyber threat intelligence production — the intelligence cycle, IOC extraction/normalization/enrichment, STIX/TAXII and MISP, structured analytic models (Diamond, Kill Chain, ATT&CK), source scoring, actor/campaign tracking, and finished intelligence reporting.
ads-performance-analytics
How to read paid media dashboards without fooling yourself. Attribution models, platform reporting quirks, multi-platform reconciliation, ROAS vs LTV horizon traps, statistical noise in performance metrics, incrementality testing, and the failure modes that produce expensive lessons. Triggers on read paid media…
ads-creative-development
How to produce ad creative that converts at performance scale. Hook patterns, format selection, video pacing, variation systems, sequential testing methodology, fatigue detection, brand-voice alignment without conversion dilution, and platform-specific creative norms. Triggers on ad creative, ad design, hook patterns…
paid-media-strategy
A discipline for running paid media that does not light money on fire. Hypothesis writing for paid spend, channel selection, budget allocation, audience targeting, bid strategy, campaign types, what NOT to spend on, attribution reality, and the failure modes that produce expensive lessons. Triggers on paid media…
ecommerce-full-pipeline
An e-commerce workflow assistant covering product discovery, 1688 product collection, listing on multiple platforms, promotional copy, and short-video creation. 1688 is a Chinese online wholesale marketplace.
absolute-simplify
Use when the user wants to simplify, clean up, refactor, tidy, or refine code — their staged/unstaged git changes or a target file/path. Reduces complexity, flattens nesting, removes redundancy and dead code, scores each change by value (holding low-value churn), then runs tests to prove nothing broke. Invoke on…