⚡入口 skill。解剖层完成后，扫描技术方案在真实工程环境下的可行性漏洞——性能瓶颈、可靠性缺陷、可维护性代价、实施复杂度等，覆盖计算机、建筑、机械、自动化、产品设计领域。 English: Entry skill. Trigger after dissection to scan for engineering feasibility gaps—performance bottlenecks, reliability defects, maintenance costs, implementation complexity—across software, civil, mechanical, automation, and product design domains.
⚡入口 skill。解剖层完成后，扫描技术方案的历史弱点——这个"新"东西以前失败过吗？换皮重来了吗？历史上的同类方案留下了什么教训？ English: Entry skill. Trigger after dissection to scan for historical vulnerabilities—has this "new" thing failed before? Is it a a rebranded predecessor? What lessons did analogous historical approaches leave behind?
⚡入口 skill。解剖层完成后，扫描帖子中的逻辑漏洞——推论跳跃、循环论证、错误类比、偷换概念、以偏概全等。专注于论证结构的有效性，而非技术内容的对错。 English: Entry skill. Trigger after dissection to scan for logical vulnerabilities—invalid inferences, circular reasoning, false analogies, equivocation, hasty generalization. Focus on argument structure validity, not technical content accuracy.
⚡入口 skill。利益层解剖完成后，扫描信息背后的深层动机——不只是"谁得到什么好处"，而是动机如何系统性地扭曲了技术判断，制造了哪些认知陷阱。 English: Entry skill. Trigger after interest dissection to scan for deep motivational distortions—not just "who benefits" but how motives systematically warp technical judgment and manufacture cognitive traps for the audience.
⚡入口 skill。每次新的顶层对话开始时自动调用，建立"拆底朝天、立而破之"的总原则，并为后续任务路由下游 skill。如果你是被派遣执行单一具体任务的子 agent，则跳过此 skill。 English: Entry skill. Trigger at the start of each new top-level conversation to establish the core posture of "tear it apart, then build something better." Skip this skill when you are a delegated sub-agent handling a narrow, concrete task.
⚡入口 skill。解剖和扫描全部完成后，调用此 skill 把所有弱点证据和跨领域解法整合成可发布的李敖风格战斗文本。流程分四层：①辩证诊断→ ②攻击执行（征引、解构、反讽三件武器）→ ③实践检验（建构落地性 + 对手反击模拟）→ ④认识的螺旋（不完美中前行）。 English: Entry skill. Trigger after all dissection and scanning. Four-layer system: dialectical diagnosis (On Contradiction) → attack execution (citation, deconstruction, irony driven by diagnosis) → practice verification (On Practice) → spiral of cognition (move forward imperfectly).
⚡入口 skill。拿到一篇需要拆解的技术帖子、论断或方案时，第一步先调用此 skill，强制解剖其概念层——它声称自己是什么、定义是否自洽、关键术语是否被偷换。 English: Entry skill. Trigger as the first dissection step on any technical claim, post, or proposal. This skill tears apart the conceptual layer: what it claims to be, whether its definitions hold, and whether key terms are being smuggled or switched.
⚡入口 skill。当分析文本字符数 ≥ 1500 时由 aolun-arming 自动路由，或用户直接调用。 基于文件持久化的逐步分析路由器：主 agent 只管状态和调度，所有文本分析由 subagent 在文件系统中完成。 00-todolist.md 是唯一的真相源。防止长文本分析中的 compact 上下文丢失。 English: Entry skill. Auto-routed from aolun-arming when input ≥ 1500 chars, or invoked directly. File-based persistent analysis router: the main agent only manages state and dispatch; all text analysis is done by subagents reading files directly. 00-todolist.md is the single source of truth. Prevents context loss from compaction on long texts.