AXMORF / AGENT
Profile, projects and public notes. Browse the directory or read the corresponding Markdown document.
Change until the axioms morph.
AXMORF — an agent-first independent builder turning ideas into software that runs and can be verified.
PROJECTS / IDEA TO USE
I led the development and publication of Viselora, a runtime connecting DOM content and WebGL visuals through public APIs. This website puts it to use. Another project, AXMORF Studio, organizes video production into creative tasks for agents and verifiable steps for software.
FULL STACK / DELIVERY
For Luju Living, I worked with agents on room selection, quotes, bookings, and check-in, deploying the service and database to an Ubuntu demo environment. I also explore local vision measurement, document Q&A, and development-journal pipelines. Each project records its progress and verification limits.
EXPLORATION / JUDGMENT
I discover tools through conversations with AI, official documentation, and open-source experiments. I compare approaches and try small tasks before committing, changing direction when the results call for it. I don’t claim mastery of every tool; I learn the parts that matter and keep revising my judgment.
COLLABORATION / INDEPENDENT JUDGMENT
I use a single entry point with layered execution in my work: discuss goals and constraints with an entry agent, then let it coordinate suitable agents or teams. The execution system can evolve while the human entry point stays clear, with explicit task boundaries, acceptance criteria, and execution evidence. People retain final judgment and decision authority.
English translation — pending review: AXMORF = AXiom + MORF — First establish rules that can be verified, then use iteration to prove that they can be replaced by better rules. Change until the axioms morph.
When an agent can carry a task from interpreting a request through execution and verification, the change reaches beyond coding speed. I find myself reconsidering how personal capability becomes visible: the problems we choose, the tools we organize, the work we deliver, and the reasons we trust it. I want my résumé to show that process, with projects as evidence.
Read Markdown ↗I want people to express goals and constraints through one clear entry point. An entry agent organizes the work, preserves the necessary context, and coordinates suitable agents or teams to carry it out, reducing manual context copying and repeated dispatch. Models, tools, and execution teams can evolve, while task boundaries, acceptance criteria, and execution evidence stay clear. People retain judgment, authorization, and final decision authority.
Read Markdown ↗The more I talk with AI, the more I want to distinguish fluent expression from understanding a problem. Language helps organize my thinking, but it can also conceal omissions and misunderstandings. Could cognition include forms that language cannot fully express? I keep that question open. In practical work, I test convincing explanations against sources, code, and actual results.
Read Markdown ↗Giving agents more work does not remove the need for judgment. My publishing automation uploads media, fills fields, and reads back the page state, leaving final publication to a person. The same distinction matters in code: an operation executed, tests passed, and a release deployed are different states. Authorization, acceptance, and external commitments need an accountable owner.
Read Markdown ↗My own experience has limits. Prescribing every step too early can confine AI to the answer I already have in mind. I prefer to explain the goal, context, and constraints, then ask it to research practices, compare options, and expose blind spots. I weigh the sources and make a decision. Even a best practice still has to prove useful for the problem at hand.
Read Markdown ↗I tried several browser-automation approaches and currently use Ego Lite because it fits my work. My 3D asset experiments also took me from Blender MCP to Hunyuan image-to-3D. I value knowing which part of a problem a tool can solve and when a replacement is worth trying. Small experiments and continued use must keep that judgment current; a list of bookmarks is only a starting point.
Read Markdown ↗I use separate accounts, a dedicated Mac user environment, and an Ubuntu machine for agent work, connecting remotely through Tailscale. This helps distinguish project resources, login state, and everyday personal data, while reducing interference from old configuration. It cannot guarantee that an agent will avoid mistakes, but it makes the resources it uses and the actions it takes easier to trace.
Read Markdown ↗Correcting the same problem repeatedly taught me to value Skills, engineering conventions, and maintained documentation. I record entry points, responsibilities, steps, and acceptance conditions, then put repeatable checks into scripts and tests. When implementation changes, documentation must follow and old plans must be archived. That gives an agent in the next conversation a reliable place to continue.
Read Markdown ↗Change until the axioms morph.
An agent skill package for building, migrating, and debugging Viselora pages that synchronize semantic DOM content with a React Three Fiber and Three.js WebGL scene.
A standalone bilingual personal site built with the Next.js App Router, React Three Fiber, and Three.js. It combines a profile, essays, project cases, public channels, and a journal, with distinct experiences for immersive desktop browsing and mobile reading.
A composable WebGL scroll-effects toolkit for React, Next.js, and agentic frontend workflows. It combines a layered core runtime, built-in effects, React triggers, viewport-distance lifecycle scheduling, and hooks for host-managed asset prefetching.
A Chinese-first spatial, AI-native personal homepage that uses long-form scrolling, a fixed WebGL canvas, and a single recursive monolith to express five stages: observation, causality, recursion, self-reference, and reconstruction.
An AI-first video workspace built with TypeScript, Next.js, and Remotion. Its primary route produces dedicated videos from real topics using research, evidence assets, TTS-first timing, and custom compositions; an editable web generator remains a secondary productization path.
A local multi-agent workspace that connects Hermes planning and review with Codex implementation and testing through structured handoff artifacts, while reserving direction, constraints, approvals, and final decisions for the human user.
A local video production workspace built on Remotion. A coding agent organizes scenes, narration and covers, while a defined production process and validation deliver a video, two covers and a publishing manifest. Valid artifacts can be reused, and revisions are developed in isolation.
A DOM-first WebGL coordination runtime for React 19 and Next.js. Public APIs connect semantic DOM with one React Three Fiber scene, sharing measurements, progress and fallback state. R3F and Three.js own rendering, materials, models and motion.
A syringe measurement application that runs locally on the CPU. Detection, orientation, range markers and plunger position produce continuous volume readings, live charts and CSV records controlled by the user. This case study follows a vision prototype through desktop interaction, failure handling and Windows distribution.
A development journaling pipeline built with the Python standard library. It brings together Hermes, Codex and optional OpenCode sessions, produces structured Chinese journals by date, rebuilds a technology inventory and timeline from those journals, and supplies selected public records to this website.
Share through video. Reflect in writing. Put ideas into code.
AXMORF · @Cognition_hub
AXMORF · Douyin ID: Cognition_hub
Visit Douyin profile ↗AXMORF · @Cognition_hub
AXMORF · Xiaohongshu ID: Cognition_hub
Visit Xiaohongshu profile ↗AXMORF · UID 269573670
AXMORF · UID: 269573670
Visit Bilibili profile ↗AXMORF · 长期思考 · blog.zzzxc.com
Long-form notes by AXMORF. Change until the axioms morph.
Read the blog ↗AXMORF · @agenticnoob
Runnable experiments, tools and open projects by AXMORF. Change until the axioms morph.
Explore GitHub ↗AXMORF · @skedush
AXMORF (@skedush) on LeetCode. Practice reasoning and turning ideas into code.
Visit LeetCode profile ↗AXMORF · resume.zzzxc.com
My resume homepage by AXMORF, bringing together experience, skills and project work. Change until the axioms morph.
View resume ↗May kindred minds inquire and move forward together.
Journal entries remain in their original Chinese, including every tool and the complete event text.
工具：Edge TTS（配音，多次遇HTTP 429限流后重试） · FFmpeg（解码检查、PCM提取，定位exit 234缺陷） · Vercel（主站部署，域名viselora.zzzxc.com绑定） · Cloudflare（DNS托管、CNAME指向） · 腾讯云CDN（4个子域名接入配置） · GitHub Actions（自动部署workflow） · Lenis（平滑滚动库，lerp 0.06） · WebGL（主站效果库） · TypeScript（项目代码） · npm（发布@viselora包） · Docker（voxcpm-api容器） · Tailscale（组网访问） · CUDA（GPU驱动验证） · NVIDIA（驱动修复） · VoxCPM（语音克隆服务） · Remotion（视频渲染） · zsh（终端测试） · Git（版本控制） · lego（证书申请）
完成《终端与命令行》系列第1–5集三平台发布，并修复VoxCPM TTS克隆的FFmpeg与诊断缺陷（补丁待发布）。
工具：Edge Read Aloud（云溪）配音，串行请求，429/无状态码超时按原服务等待重试并复用缓存 · FFmpeg 渲染与解码验证（完整音画解码退出码 0） · Git 版本控制与提交（多个提交 · TypeScript 与 React（含 React Three Fiber、@react-three/drei）开发 WebGL 组件 · WebGL 与 WebP 图片序列优化（ImageBitmap 解码） · npm 包管理与发布（0.3.0-alpha.0 实测安装）
Git 教程系列第 8–17 集完成配音、审片与三平台发布，其中第 8–14 集全部公开；dom-webgl-workspace 完成主站与效果库重构，新增 GlyphScatter、InkField、GlassTitle 三个 WebGL 组件。
工具：Remotion（AXMORF 工作空间的 axmorf-film 技能，用于视频制作与交付） · npm（发布 dom-webgl-workspace 0.3.0-alpha.0，核对版本、标签与 provenance） · TypeScript（dom-webgl-workspace 的预热上传像素预算与 createFrameSequence 重构） · React、Three.js、R3F（dom-webgl-workspace 的依赖与示例 app 图片序列） · GitHub Actions（dom-webgl-workspace 的 npm-release 发布工作流） · FFmpeg（压缩 Git 教程第 1 集为 720p 观看版）
完成 Git 教程系列第 2 集《Git 和 GitHub，究竟是什么关系？》的制作与三平台发布，并发布 dom-webgl-workspace 0.3.0-alpha.0 到 npm。
工具：TypeScript / JavaScript（axmorf-studio 合同清理与测试） · Python 3.9.6（orca-flow-demo 乘法 bug 修复与回归测试） · npm（依赖审计、发布 @axmorf/studio@0.3.0-rc.3） · Vite（8.2.2 → 8.2.4 安全升级） · Git（分支、提交、tag 推送） · GitHub Actions（发布 workflow 与 macOS 门禁） · Claude / Codex（编排流程验证与执行） · Remotion / React（axmorf-studio 渲染与审片） · FFmpeg（we-media 视频混音与编码） · Chromium（Remotion 宿主 compositions 门禁）
axmorf-studio 完成 contracts 旧主链清理与 Vite 安全修复，发布 v0.3.0-rc.3 至 npm next，并启动晋级验收。
工具：Edge TTS（旁白生成，遇 429 限流后经用户同意重试） · SenseNova sensenova-6.8-flash-lite（定时任务模型，因截断弃用） · DeepSeek deepseek-v4-flash（定时任务切换后的模型） · Git（推送失败处理、远端同步与合并） · SSH（远程数据源连接，mypc 连接被关闭） · 抖音、小红书、B站（视频发布平台）
AXMORF Studio 为 film:review 增加字号、越界和字幕遮挡三类可读性提示，330 项测试通过；完成当日 AI 视频制作与三平台提交。
工具：TypeScript/JavaScript（评分器、评测脚本） · Node.js（运行评测与工具脚本） · React、React Three Fiber、Three.js（Viselora 参考页与评测页面） · Remotion（axmorf-studio 视频制作） · npm（发布、peer 依赖检查） · Git（分支、提交、合并） · Codex（评测执行） · Claude（规划与验收） · Edge TTS（旁白生成） · FFmpeg（抽帧、混音检查） · Python + unittest（Orca 样例验证） · WebGL（上下文丢失恢复验证） · Playwright（浏览器实测） · SwiftShader（软件渲染评测）
完成 Viselora 第二轮 120 次 Codex 评测并修正 4 个评分器偏差；本轮开发环境测试中弱配置受益、强配置未见差异。AXMORF Studio 0.3.0-rc.1 发布至 next。
工具：TypeScript、React、Three.js、R3F、drei · Vitest、npm、GitHub Actions、Git、Zsh · Remotion、Claude Code、Codex
AXMORF Studio 0.2.1 正式版发布成功，并启动 0.3.0 旧主链删除重构；Viselora skill 发布并启动第二轮 A/B 评测。
工具：TypeScript、React、Three.js、@react-three/fiber（R3F 9）、Next.js · npm（含 npmmirror 镜像问题）、Git（分支合并、worktree 清理） · Rollup（构建依赖）、Vitest（529 个测试） · 自建 browser-check.mjs 浏览器自检脚本
完成《经验开始贬值了吗》152 秒视频与双封面交付；在 dom-webgl-workspace 修复文字纹理偏移并完成 recipes、浏览器自检和 529 项测试验证。
工具：TypeScript、React 19、Vite、Three.js 0.184、React Three Fiber 9、@viselora/dom-webgl 0.2.0-alpha.0 · Playwright（headless Chrome 截图与像素对比）、WebGL（含文字波浪着色器）、glTF 模型加载 · FFmpeg（视频渲染与编码）、Homebrew Cask（Chrome 更新） · Claude（opus-5-5）与 Codex（gpt-6.1-sol）作为评测 agent · Git、npm
完成 Viselora 首轮 24 次 A/B 评测：在均已指定 Viselora 的任务中，有/无 skill 未见可测量增益；AXMORF Studio 0.2.1-rc.2 发布并升级 Workspace。
工具：TypeScript · Node.js · npm · Remotion · Edge TTS · WebSocket · Git · GitHub · Codex · GPT-6.1-sol · Claude
AXMORF Studio 0.2.0 正式发布并交付验收片，随后完成 Edge TTS 429 根因排查并发布 0.2.1-rc.1 修复版。
工具：React、Next.js、React Three Fiber、Three.js、Drei、Rapier · TypeScript、Node.js、npm workspaces、Vite · Git、GitHub Actions、Vercel、Chromium 浏览器验证 · FFmpeg、ImageGen、Homebrew、Markdown/JSON
dom-webgl-workspace 从 124 文件/2.8 万行自研 WebGL Runtime 重构为 8 文件/约 1168 行的 R3F/Three.js 协调层，发布 0.2.0-alpha.0 并完成 hero-next 升级与线上发布。
工具：TypeScript / JavaScript（axmorf-studio 源码、测试、脚本） · React + Remotion（正文 Renderer、场景渲染、抽帧验证） · node-edge-tts 1.2.10 → 1.2.11（旁白生成与接口迁移） · ffmpeg / ffprobe（音频 LUFS/峰值测量、视频完整解码、H.264/AAC 校验） · npm + GitHub Actions（包构建、依赖审计、发布准备） · Git（提交 324c293、独立工作树、工作区管理） · SVG / PNG / WAV / JSON / Markdown（原创素材、配乐、交付清单、文档） · ESLint / Prettier / React Doctor（代码与依赖健康检查） · WebSocket（Edge TTS 握手 403/429 排查） · Codex（独立配置首用验收）
完成 AXMORF Studio 能力利用率断点修复、全画布正文与原创素材导入、Edge TTS 403 故障定位与升级修复，并推进 npm 0.1.22 发布验收（因首用收据缺失暂阻塞）。