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 hollandkevint/data-product-operator --skill arbitrage-audit-datagit clone --depth 1 https://github.com/hollandkevint/data-product-operatorWrote 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/hollandkevint/data-product-operator/arbitrage-audit-data)<a href="https://agentmods.dev/skills/hollandkevint/data-product-operator/arbitrage-audit-data"><img src="https://agentmods.dev/badge/skills/hollandkevint/data-product-operator/arbitrage-audit-data/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/hollandkevint/data-product-operator/arbitrage-audit-data"><img src="https://agentmods.dev/badge/skills/hollandkevint/data-product-operator/arbitrage-audit-data.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 | $0.00066 | $0.00788 |
| Opus 5 | $0.00033 | $0.00394 |
| Sonnet 5 | $0.00013 | $0.00158 |
| Haiku 4.5 | $0.00007 | $0.00079 |
Grade A, and why
arbitrage-audit-data 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 9d 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 — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The 3 Questions (Data Product Context)
Question 1: What inefficiency is this data product built on?
Every data product sits on top of a gap. Name it:
| Gap Type | Data Product Example | Closing Speed |
|---|---|---|
| Knowledge asymmetry | "Only our analysts know how to calculate this metric" | Fast — AI can learn metric definitions |
| Fragmentation | "Data lives in 5 systems nobody has connected" | Medium — integration tools accelerating |
| Speed | "This report takes 3 days of manual SQL" | Fast — automation and agents |
| Discipline | "People skip the quality checks" | Medium — AI can enforce process |
| Judgment | "Someone needs to decide which cohort definition is clinically valid" | Slow — requires domain expertise |
| Relationship | "The client trusts our interpretation, not just the numbers" | Slow — fundamentally human |
Ask: "If a competitor had the same data and unlimited AI, what would still be hard for them to replicate?"
Question 2: How fast can AI close this gap?
Informational data products (reports, dashboards, automated queries) face fast closure. If your data product's value is "we run the SQL so you don't have to," the clock is ticking.
Judgment data products (cohort validation, clinical interpretation, business context) face slow closure. If your data product's value is "we know what this number means for YOUR situation," that's durable.
Specific data product signals:
| Signal | Gap Closing | Action |
|---|---|---|
| Consumer could get the same answer from ChatGPT + raw data | Fast | Migrate to judgment layer |
| Consumer needs your domain expertise to interpret results | Slow | Encode and protect that expertise |
| Consumer uses your output as input to another automated system | Fast | The consuming system will eventually skip you |
| Consumer uses your output to make human decisions | Slow | Double down on decision context |
Question 3: What new gap opens when this one closes?
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.
- 9d ago First seen · 79 lines · 66 tokens per session scan A fda60948b29e
arbitrage-audit-data is a skill published in the GitHub repository hollandkevint/data-product-operator (3 stars, last pushed today), licensed MIT. It adds 66 tokens to every session and 788 once invoked, about $0.0003 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
fechar-dia
Fecha o dia de trabalho. Avalia as 3 do dia, classifica o que não fechou, semeia o movimento.
preparar-conversa
Monta a pauta antes de uma conversa que importa. 1:1 com liderado, 1:1 com o próprio líder, alinhamento com par ou stakeholder. Use quando a pessoa disser "prepara o 1:1 com [nome]", "monta a pauta", "vou falar com [nome]", "tenho reunião com [nome], me prepara", "/preparar-conversa". NÃO use pra processar a conversa…
processar-reuniao
Transforma uma reunião que aconteceu em nota estruturada e propaga o que importa. A pessoa cola a transcrição, o resumo automático ou as notas cruas. Use quando a pessoa disser "processa essa reunião", "processa esse 1:1", "olha o que rolou na reunião", colar uma transcrição pedindo organização, ou…
classify
Assess or re-assess the innovation-vs-implementation classification and 0-100 innovation weight for one feature or the whole Product/Features set, then regenerate the ranked features-index.md and run the portfolio sanity check. Use when features need (re)ranking or after editing the UVP.
feature
Create or revise a single Product feature file with a complete user story, Given/When/Then acceptance criteria, and an innovation-vs-implementation classification. Use when adding one feature to Product/Features/, or refining an existing feature, outside a full /innovate run.
iflow
The front door / orchestrator for the I-Framework. Reads the Product/ folder, determines which stage (Intent, Ideation, Innovation) you're in from the artifacts and their gates, reports the risk-burndown state, and routes you to the next step. Use to start or resume the Intent → Ideation → Innovation flow.