Most macro analysts already maintain some version of a country file. It may live across notebooks, spreadsheets, saved links, charts, emails, call notes, and half-remembered data caveats. The problem is not that the knowledge does not exist. The problem is that it is fragmented, hard to reuse, and often invisible to both colleagues and AI tools.

A per-country AI skill offers a more structured alternative. Instead of treating a skill as a generic prompt template, the analyst can use it as a living, country-specific operating file. For each country, the skill captures the analyst’s current view, scenario probabilities, source map, preferred data series, known comparability problems, and relevant network contacts.

The structure can be simple:

• current-view.md holds the baseline, upside, and downside scenarios, each with probabilities and signposts.
• sources.md records trusted primary and secondary sources.
• data-map.md explains which data series to use for inflation, fiscal policy, debt, labor markets, external balances, and growth, including CountryData.io indicator codes where useful.
• analytical-playbook.md captures the country-specific mental model: recurring traps, transmission channels, and the key questions to ask before forming a view.
• contacts.md records who to call for what.

This is different from building a country encyclopedia. The goal is not to reproduce public knowledge about Switzerland, Ghana, Brazil, or Indonesia. The goal is to capture the analyst’s judgment: what matters, what is misleading, which data series are trusted, which indicators are not comparable across countries, and what evidence would change the view.

There are already related efforts in AI-enabled finance. Financial-services skills are being built for modeling, equity research, earnings analysis, and investment memos. Agentic investment-research platforms can gather data, summarize filings, and produce reports. Macro tools increasingly generate base, bull, and bear scenarios around data releases. But the per-country analyst skill is a slightly different object. It is not just automation. It is institutionalized judgment.

That distinction matters for incentives. In a corporate setting, analysts may worry that encoding their knowledge into a skill weakens their position. But the opposite can be true if the skill is understood correctly. A detailed country skill makes expertise visible. It shows the depth of the analyst’s source network, data judgment, country intuition, and scenario discipline.

It also makes clear that the skill is not self-maintaining. Without the expert, the file depreciates. Views become stale, contacts go cold, institutions change, indicators break, and old caveats stop applying.

In that sense, the skill increases the analyst’s bargaining power. It documents the value of the expert while making the expert more productive. The company gets a more reusable analytical asset. The analyst gets a clearer record of accumulated country knowledge and a stronger claim on the ongoing maintenance, interpretation, and updating of that knowledge.

This level of transparency also raises the bar for the analyst. Once the country skill makes the analyst’s assumptions, scenarios, source hierarchy, data choices, and prior judgments explicit, it becomes much harder to hide behind generic commentary or recycled market consensus.

The file reveals whether the analyst is adding differentiated insight or merely summarizing what is already public. It therefore creates pressure to be precise: to explain what is genuinely known, what is uncertain, where the analyst disagrees with consensus, and what evidence would force a change in view.

In that sense, the skill is not only a productivity tool. It is also a discipline mechanism. It rewards analysts who maintain a real edge through judgment, source work, and data interpretation, while exposing analysis that is mostly repetition.

The best version of this approach is therefore not “replace the analyst with a skill.” It is: make the analyst’s country expertise legible, reusable, and harder to undervalue.