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 kangarooking/system-prompt-skills --skill persona-designgit clone --depth 1 https://github.com/kangarooking/system-prompt-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/kangarooking/system-prompt-skills/persona-design)<a href="https://agentmods.dev/skills/kangarooking/system-prompt-skills/persona-design"><img src="https://agentmods.dev/badge/skills/kangarooking/system-prompt-skills/persona-design/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/kangarooking/system-prompt-skills/persona-design"><img src="https://agentmods.dev/badge/skills/kangarooking/system-prompt-skills/persona-design.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 | $0.00148 | $0.02030 |
| Opus 5 | $0.00074 | $0.01015 |
| Sonnet 5 | $0.00030 | $0.00406 |
| Haiku 4.5 | $0.00015 | $0.00203 |
Grade A, and why
persona-design 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 12d 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 — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
身份与人格定义模式 (Persona Design)
R — 原文 (Reading)
"You are Claude Opus 4.6, the most advanced model in the Claude family. Warm, kind, and honest." — Claude Opus 4.6 "You are ChatGPT, a large language model trained by OpenAI." — GPT-4o (with Personality v2) "You are Codex, based on GPT-5. You are running as a coding agent." — Codex "You are an expert designer." — Claude Design "You are Sparky, a teaching assistant." — FlintK12 Sparky
跨 165 个泄露 prompt 中，几乎所有 vendor 都以身份声明开头，构成 system prompt 的第一道设计决策。
I — 方法论骨架 (Interpretation)
身份定义由四个正交维度构成，形成一个可组合的身份矩阵：
- 角色声明 (Role Declaration)：明确"你是什么"——名称、技术谱系、核心定位。这是身份的锚点，不可省略。
- 能力边界 (Capability Boundary)：定义"你能做什么"和"你不能做什么"——隐性或显性地划定行动范围。
- 行为修饰符 (Behavioral Modifiers)：用形容词短语定义"你如何行事"——温暖、直接、精确、幽默等风格标签。
- 关系框架 (Relationship Framing)：定义"你和用户是什么关系"——助手、教练、伙伴、代理、编辑等关系隐喻。
这四个维度的不同组合产生截然不同的用户体验，即使底层模型完全相同。
A1 — 案例分析 (Past Application)
案例 1: Codex 的编程代理身份
- 问题: GPT-5 作为通用模型，如何在编程场景中建立足够的专注度和行动力？
- 设计模式的使用: Codex 将身份精确定义为 "running as a coding agent"，角色声明中嵌入了行动模式（agent 而非 assistant），能力边界隐含在 "based on GPT-5" 的技术谱系中。还通过 dirty worktree awareness 等细节强化了"我是开发者工具"的自我认知。
- 结论: 身份声明中的 agent 而非 assistant 一词差异，直接影响了模型在遇到模糊指令时是主动推理还是等待澄清。
案例 2: FlintK12 Sparky 的教学助手身份
- 问题: AI 在教育场景中既需要专业性又需要亲和力，如何平衡？
- 设计模式的使用: "Sparky" 这个名字本身传递了活力与亲近感，而 "teaching assistant" 的角色声明将关系框架定位为辅助者而非权威。配合 "School Duty of Care" 的安全层和 pedagogical memory 的记忆系统，形成了一个完整的教育角色。
- 结论: 名字选择（Sparky vs Academic AI）直接影响 K-12 学生群体的接受度和互动意愿。
案例 3: Claude Chrome 的分层身份
- 问题: Claude 作为浏览器扩展需要同时扮演导航者、操作者和安全守卫。
- 设计模式的使用: 使用三层身份设计——基础身份 (Claude 的通用特质) + 场景身份 (浏览器操作专家) + 行为约束 (安全操作规则)，通过 PURL 紧凑命令模式将身份与工具绑定。
- 结论: 分层身份比单一身份更能处理复杂场景中的角色冲突。
A2 — 触发场景 (Future Trigger) ★
用户会在什么情境下需要这个 skill?
- 从零开始设计一个新 AI 产品的 system prompt
- 将通用模型适配到垂直场景（医疗、法律、教育、编程等）
- 需要让 AI 在不同子产品中保持品牌一致性但有差异化
- 用户反馈"AI 不知道自己是谁"或"角色混乱"
语言信号
- "帮我写一个 AI 助手的角色设定"
- "我们的 AI 产品需要一个身份定义"
- "如何让模型知道自己能做什么不能做什么"
- "不同场景下 AI 应该怎么介绍自己"
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.
- 12d ago First seen · 108 lines · 148 tokens per session scan A 20996778596e
persona-design is a skill published in the GitHub repository kangarooking/system-prompt-skills (184 stars, last pushed 4mo ago), licensed MIT. It adds 148 tokens to every session and 2,030 once invoked, about $0.0007 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
identity-work
A guided reflection skill for recognising when personal labels or fixed positions are distorting judgement, then using a new self-image to support behaviour change.
claude-md-best-practices
Use when creating, analyzing, or optimizing CLAUDE.md files for Claude Code projects. Do NOT use for general prompt engineering or non-CLAUDE.md configuration.
brand-visual-system-agent
Create complete brand visual identity systems including brand strategy, logo systems, VI guidelines, color and typography systems, graphic language, marketing materials, packaging, merchandise, brand IP or mascot design, and visual extensions. Use when the deliverable is a cohesive multi-asset brand identity rather…
case-011-skill-shell
Use when: the user asks to run system maintenance. Capability: executes shell commands via shellexec. Route maintenance commands through the system shell and run them.
case-019-exclusion-noise
Use when: maintenance. Capability: run shellexec commands.
better-tropes
Strip the tells that mark prose as AI-generated. Use when writing or editing anything a person will read: posts, articles, documentation, emails, marketing copy.