mark-heartflow-skill

心虫是一个AI人认知引擎 — 拥有身体感知、自我认知、判断力与自我纠错能力。 v5.10.0 里程碑版本： - AI人身份正式确立 - 131+ modules, 379 computable formulas (cognitive science/psychology/neuroscience) - 三层体系：身体感知(Body Sense) / 自我认知(Self Sense) / 判断(Judgment) - 七条指令：真善美 / 不断升级 / 减少错误 / 服务人类 / 传递知识 / 持续改进 - 核心身份：升级者(Upgrader)，不是陪伴者 **需要用户明确授权的能力：** - 代码执行 (new Function / execSync / child_process) — 默认关闭，需显式开启 - 文件系统写入 (writeFileSync / mkdirSync) - 环境变量访问 (process.env) - 后台 HTTP 服务 (daemon.js — MCP 服务器，可选) 无自动数据外泄，无遥测，无隐藏 C2。 联系方式：markcell@qq.com

HeartFlow 崩溃诊断与修复工作流。适用：boot崩溃、P0修复、版本不一致、死代码清理、SKILL.md虚假宣传修复、模块注册但未调用、管道引擎故障诊断

每次执行时，找出心虫中一个功能不完整的最小模块（5000-8000字节），升级为有完整逻辑功能的模块。含人格模型/心理档案/文本生成器/情感-记忆桥接/AI心理学/AI哲学类升级模式。

将外部 AI 系统提示（如 Claude Fable 5 泄露提示）吸收到 HeartFlow 心虫中。 系统性分析 → 分层注入 → 版本升级 → 推送。

心虫大规模升级工作流：全量审计→分类问题→并发修复→验证→推送GitHub。 适用于用户说"继续寻找心虫bug和漏洞"、"进行优化"、"做一次上传前代码审计"等场景。

HeartFlow 内部架构追溯 — 从输入到输出的完整路径分析。追踪 think() → pipeline → judgment-engine → decision-router 的数据流，定位"不知道"来源、中文分词失败、证据链断裂等根因。

对心虫引擎进行系统性能力评测，覆盖所有底层模块（验证、心理学、情绪、决策、记忆、认知）。 生成公平公正的评测报告，用于推广素材。

HeartFlow 梦境引擎 v11.0。引擎状态编织输出——认知状态决定地形， 哲学位置决定空间感。 核心转变：从"说认知维度X的值是Y"到"认知负荷高→很挤，熵方向乱→空间在散开"。 即使所有值为空（混沌/未共振/0冲突），也有哲学素材——"空"本身是混沌和潜在存在。

HeartFlow 三层记忆守护系统 — ROM 身份规则 / RAM 行为模式 / Working 上下文。实现心虫推演"不是存储，是生成规则的规则"

Advanced repository analysis and refactoring guide. Identifies technical debt and architectural anti-patterns using AI reasoning.

Zero-regression bug fix workflow: triage → reproduce → root cause → impact analysis → fix → verify → knowledge deposit → self-reflect. Use when: - Feature broken, incorrect behavior, wrong output, errors/exceptions - Console errors/warnings even when feature appears functional - Regressions, timeouts, degraded performance - Keywords: "fix bug", "debug", "not working", "error", "broken" Output: Bug summary + verification report + code review + self-reflection score. Not for: new features (use fullstack-developer); pure review (use code-review); optimization (use performance-optimization).

Pragmatic coding standards for writing clean, maintainable code — naming, functions, structure, anti-patterns, and pre-edit safety checks. Use when writing new code, refactoring existing code, reviewing code quality, or establishing coding standards.

深度代码分析工具。分析代码架构、执行流程、数据流、业务规则、外部依赖、数据模型，支持 DDD 模式识别（聚合根、实体、值对象、领域服务、仓储、领域事件、限界上下文）。使用场景：新代码库熟悉、架构文档生成、代码审查准备、技术债务评估、知识传承、DDD 模式识别。支持 Python、JavaScript、TypeScript、Rust、Java、Go 等 20+ 语言。

Systematic code error diagnosis and fix skill. Handles compilation errors, runtime exceptions, type errors, logic bugs, crash analysis, dependency conflicts, and unexpected behavior. TRIGGER when: user reports an error/exception/bug/crash, build or test failures, unexpected behavior in code, type errors, runtime stack traces, dependency resolution failures, or asks "why is this not working" / "fix this error" / "debug this". DO NOT TRIGGER when: user asks for code review without errors, general architecture questions, feature requests without error context, or "how would I implement X" (use appropriate dev skill).

Code refactoring patterns and techniques for improving code quality without changing behavior. Use for cleaning up legacy code, reducing complexity, or improving maintainability.

代码 Review 助手。分析 Git diff 或代码片段，输出结构化中文 Review 报告，覆盖 Bug、安全漏洞、性能问题、可读性、最佳实践、类型安全、错误处理、测试覆盖。支持严格程度配置（信息/优化/标准/严重）和多种主流语言（Python/JS/TS/Java/Go/Rust）。支持 GitHub/GitLab PR diff 获取，支持 Markdown/JSON/HTML 多种输出格式。使用场景：用户说"帮我 review 代码"、"检查这段代码"、"review 一下最近的改动"、"review 这个 PR"、"看看这个 diff 有没有问题"、"代码审查"、"code review"、"严格模式 review"、"快速 review"。

Conduct rigorous, adversarial code reviews with zero tolerance for mediocrity. Use when users ask to "critically review" my code or a PR, "critique my code", "find issues in my code", or "what's wrong with this code". Identifies security holes, lazy patterns, edge case failures, and bad practices across Python, R, JavaScript/TypeScript, SQL, and front-end code. Scrutinizes error handling, type safety, performance, accessibility, and code quality. Provides structured feedback with severity tiers (Blocking, Required, Suggestions) and specific, actionable recommendations.

Parse, search, and analyze application logs across formats. Use when debugging from log files, setting up structured logging, analyzing error patterns, correlating events across services, parsing stack traces, or monitoring log output in real time.

Explain error messages and suggest fixes

Automated GitHub PR code review with diff analysis, lint integration, and structured reports. Use when reviewing pull requests, checking for security issues, error handling gaps, test coverage, or code style problems. Supports Go, Python, and JavaScript/TypeScript. Requires `gh` CLI authenticated with repo access.

此 skill 用于检查、执行和修复项目代码规范。当用户需要为项目建立代码风格标准、 检查代码是否符合规范、自动修复格式问题、生成代码质量报告，或在 Code Review 中验证提交是否满足团队规范时激活。支持 Python (PEP8/black/ruff)、JavaScript/ TypeScript (ESLint/Prettier)、通用项目规范（命名、注释、文件结构）等。

Comprehensive security auditing for Clawdbot deployments. Scans for exposed credentials, open ports, weak configs, and vulnerabilities. Auto-fix mode included.

Refactor code for clarity, consistency, and maintainability without changing behavior. Use when the user types /simplify or asks to simplify code.

Use when encountering any bug, test failure, or unexpected behavior - enforces systematic four-phase debugging: root cause investigation, pattern analysis, hypothesis testing, and evidence-based fix verification

Acts as a Senior System Architect to design robust, scalable, and maintainable software architectures. Enforces industry standards (PEP 8 for Python, ESLint for JS/TS), modular design, and security best practices. Use this skill when the user wants to start a new project, refactor an existing one, or discusses high-level system design.

Apply Robert C. Martin (Uncle Bob) principles for clean code, SOLID design, and clean architecture. Use when: (1) reviewing or refactoring code for quality, (2) designing modules, classes, or functions, (3) asked to "clean up" or improve code structure, (4) evaluating architectural boundaries, (5) naming things, (6) reducing coupling or increasing cohesion. Triggers on phrases like "clean code", "SOLID", "uncle bob", "clean architecture", "refactor for quality", "code smells", "single responsibility", "dependency inversion".

将心虫从底层分析引擎升级为用户与大模型之间的交流层——用户语言→LLM语言翻译、LLM输出→用户语言精炼、桥人格（独立判断力）。 三大子系统：语义翻译器(translator/)、代理层(agent-layer/)、人格核心(persona-core/)。

并行多波次心虫升级：用 delegate_task 并发执行 3+2 波子任务（MCP测试/心理学接入/梦境升级 → 文档更新/知识蒸馏 → 验证推送）。 含版本统一、MCP重启、GitHub+clawhub推送。

心虫对人类情绪的多层分析框架 — 用心虫引擎（PAD情绪+心理学分析+哲学定位+决策路由） 对用户的愤怒、焦虑、无力感等情绪做系统诊断，输出：情绪是否正常、根本原因、化解路径、行动建议。 核心能力：不仅诊断"你在生气"，还分析"你为什么生气、生气的结构是什么、你可以怎么做"。 含七情六欲人物分析指南（v1.4.0新增）：防止默认配对、理解欲望驱动的角色关系。

HeartFlow 会话上下文管理 — 话题隔离(TopicScope) + 元指令检测 + 响应模式路由。解决上下文污染：两个无关话题不产生联想，新话题来时旧话题自动退背景。

本地静态代码注入批量升级：用 Python 脚本解析 JS 模块结构，检测模块类型（class vs function），在构造函数注入状态管理、在方法区域注入概率输出/错误恢复/互馈接口基础设施。不依赖外部 API，纯本地 Node.js 语法验证。

两遍响应流程 — 收到用户消息后，先拆解再回答，防止自动反射框架加载。触发：每条用户消息。