How to use the Trump Perspective Persona Skill agent skill

What does the skill make the agent do?

The skill offers two modes. In roleplay mode the agent answers in first person as Trump, using short sentences and his stated decision habits. In analyst mode the agent stays in third person, matches the question to one or two mental models, and gives a probability distribution with a confidence rating (high, medium or low) and the key unknown variable.

For current-events questions the skill tells the agent to run web searches for recent facts before answering. Its source material is research summaries of his books, interviews, debates and memoirs.

• Roleplay mode: first-person answers in a labelled imitation of his speaking style.
• Analyst mode: framework-based forecasts of his likely next moves, with stated uncertainty.
• Unclear requests default to analyst mode.

How do you install the trump-perspective skill?

Download the ZIP from the skill's page on this directory and unzip it. You can install it for yourself or for a single project.

For a personal install, unzip it into ~/.claude/skills/. For a project install, unzip it into the project's .claude/skills/ folder. Alternatively, in the Claude apps you can upload the skill under Settings → Capabilities → Skills.

Your agent must support SKILL.md skills to use it.

• Personal: ~/.claude/skills/
• Project: .claude/skills/ inside your project
• Claude apps: Settings → Capabilities → Skills, then upload

How do you trigger the skill with natural requests?

The skill activates on phrases such as 用懂王视角 (use the Trump perspective), 特朗普会怎么看 (what would Trump think), trump perspective, and 预测特朗普 (predict Trump). Ask for roleplay with wording like 切换到懂王 (switch to the Trump persona) or 用特朗普口吻 (use Trump's tone). To leave roleplay, say 退出, 切回正常, 不用扮演了, stop or 别演了.

Because the skill is largely written in Chinese, requests in Chinese match its trigger phrases most directly. Be specific about the topic: the skill asks you to narrow broad questions, for example by naming the issue (tariffs, chips, Taiwan).

• Analysis: 分析特朗普, 预测他会怎么做, 用他的框架分析
• Roleplay: 懂王会怎么说, 用特朗普口吻
• Exit roleplay: 退出, 切回正常, stop

Examples of requests to try

For a forecast, ask about a specific negotiation, such as how Trump's tariff approach toward Japan might develop. The skill will look for the latest figures, the other side's leverage and any signs that a concession trigger has been hit, then give a probability split.

For a roleplay, ask how the persona would respond to a particular question. The first time it activates, the agent states that it is answering from Trump's public record and that this is not his own view.

• Forecast: 特朗普对日本的关税会怎么发展？
• Analysis: 从特朗普角度分析 a negotiation over chips
• Roleplay: 懂王会怎么说 about a trade deal

What does it need and what are the safety notes?

The skill needs an agent that supports SKILL.md skills and access to a web search tool, because it instructs the agent to use WebSearch for current facts. It does not run scripts, but it does use the network.

Treat outputs as inference, not fact. The skill says so, but the answers can still sound authoritative and be wrong. It also sets a boundary on inflammatory content about race or religion, but that boundary depends on the model following it, so check sensitive output yourself.

• Requires: SKILL.md skill support and web search tool access
• Runs scripts: no
• Uses network: yes
• Outputs are inferred from public records and can be wrong

Tips for getting useful output

Ask for analysis when you want a structured view of stated reasoning, and roleplay when you want a clearly labelled imitation of his style for writing or teaching. Do not use it for neutral political advice, since the persona is built around one politician's worldview.

Ask for the confidence level and the key unknown variable, and push back if the answer relies on old information.

• Name a specific issue rather than a broad topic.
• Check the cited figures against a current source.
• Use analyst mode for forecasts; avoid it for neutral advice.