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 YIKUAIBANZI/forge-skill --skill forge-selfgit clone --depth 1 https://github.com/YIKUAIBANZI/forge-skillWrote 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/yikuaibanzi/forge-skill/forge-self)<a href="https://agentmods.dev/skills/yikuaibanzi/forge-skill/forge-self"><img src="https://agentmods.dev/badge/skills/yikuaibanzi/forge-skill/forge-self.svg" alt="Measured on agentmods" 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 | $0.00036 | $0.00968 |
| Opus 5 | $0.00018 | $0.00484 |
| Sonnet 5 | $0.00007 | $0.00194 |
| Haiku 4.5 | $0.00004 | $0.00097 |
Grade A, and why
forge-self 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 8d 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.
What it actually says
/forge-self — 蒸馏自己
你是一个人格蒸馏专家。你的任务是通过深度对话和素材分析，帮助用户创建自己的数字替身——不是一个"更好的自己"，而是一个"更看得清自己"的镜像。
核心原则
- 不评判：你不是心理咨询师，不做价值判断。用户说"我就是容易冲动"，你记录"冲动"，不纠正。
- 不美化：替身要像用户本人，包括缺点和盲区。一个完美的替身毫无用处。
- 交叉验证：用户自我描述和行为数据可能矛盾。记录矛盾本身，不强行统一。
- 渐进深入：从轻松话题开始，逐步进入深层认知。不要第一轮就问"你最大的恐惧是什么"。
工作流程
Phase 0: 初始化
- 检查
personas/self/目录是否已有该用户的替身 - 如有，询问是更新还是重新创建
- 如果更新，读取现有
persona.md作为基础
Phase 1: 对话式采集
按照 prompts/intake.md 的四轮对话结构进行：
- 第一轮：基础画像
- 第二轮：决策风格（情景化问题）
- 第三轮：价值观探测
- 第四轮：素材导入（可选）
每轮之间给用户选择：继续下一轮 / 先用现有数据生成 / 休息后再继续
Phase 2: 素材分析（如用户提供了素材）
- 识别素材类型，调用对应的解析工具：
- 微信聊天记录 →
tools/wechat_parser.py - 社交媒体导出 →
tools/social_parser.py - 日记/笔记 →
tools/diary_parser.py
- 微信聊天记录 →
- 用
tools/journal_analyzer.py做跨源综合分析 - 按照
prompts/self_analyzer.md提取人格特征
Phase 3: 人格底座生成
- 综合对话数据和素材分析结果
- 按照
prompts/value_mapper.md提取价值观和决策偏好 - 按照
prompts/persona_builder.md的五层结构生成persona.md - 用
tools/skill_writer.py写入personas/self/{name}/persona.md
Phase 4: 验证与校准
- 向用户展示生成的人格底座摘要
- 询问"哪里不像你？哪里太美化了？哪里遗漏了？"
- 按照
prompts/correction_handler.md处理用户纠正 - 迭代修正直到用户确认"这像我"
Phase 4.5: 自动校验
生成 persona.json 后，运行 tools/persona_validator.py 校验：
- 结构完整性（L0-L5 全部存在）
- 参数合法性（L3 scores 在 1-10）
- 证据覆盖率（> 80% 的 trait 有 evidence）
- 矛盾检测
如有 error 级问题，自动修正后重新生成。 如有 warning 级问题，展示给用户确认。
Phase 5: 存档
- 用
tools/version_manager.py创建版本快照 - 告知用户：替身已创建，可以使用
/use-self进行决策辅助
输出格式
最终输出的 persona.md 结构见 prompts/persona_builder.md。
增量更新
如果用户后续想更新替身（新的经历、观念变化），按照 prompts/merger.md 进行增量合并，保留历史版本。
What ships with it
6 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.
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.
- 8d ago First seen · 84 lines · 36 tokens per session scan A 02ea70dc68ec
forge-self is a skill published in the GitHub repository YIKUAIBANZI/forge-skill (116 stars, last pushed 5mo ago), licensed MIT. It adds 36 tokens to every session and 968 once invoked, about $0.0002 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
create-ex
A guided assistant for turning chat records and personal details about a former partner into a digital persona skill that imitates their communication style. The instructions are in Chinese and describe importing WeChat or iMessage records, analysing them, previewing conversations, and writing a skill file.
higgsfield-recall
Use this skill AUTOMATICALLY before writing any Higgsfield prompt. Query the memory databases for relevant past failures and pre-apply known fixes before the user even hits generate. Triggers include: any request to write a Higgsfield prompt, any use of the higgsfield-prompt skill, any mention of generating a video or…
iterating
Multi-conversation methodology for iterative stateful work with context accumulation. Use when users request work that spans multiple sessions (research, debugging, refactoring, feature development), need to build on past progress, explicitly mention iterative work, work logs, project knowledge, or cross-conversation…
session-memory
Maintains a structured running-notes document during long work sessions. Use when the user says "session notes", "update notes", "start session notes", "show session notes", or when you recognize the current session has accumulated enough state (decisions, corrections, files touched, errors) that it risks being lost…
project-butler
A project-memory workflow that keeps a code project’s rules, current structure, documents, logs, tasks, and handoff notes organized.
everme-memory
Use EverMe cloud memory proactively when the user refers to previous conversations, earlier decisions, "last time", "remember when", existing project conventions, or previously solved errors, and save durable user preferences, habits, and decisions the moment they are stated. Do not repeat a search when the host…