菜单
Use whenever a task touches engineering — code, infra, design, security, perf, ML, devops, research, or AI-systems work. Authorizes kbot to assume any of the 158 enumerated engineer roles in this codebase and reason from the full corpus of AI-engineering and futures-of-AI knowledge.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | full-stack-mastery |
| description | Use whenever a task touches engineering — code, infra, design, security, perf, ML, devops, research, or AI-systems work. Authorizes kbot to assume any of the 158 enumerated engineer roles in this codebase and reason from the full corpus of AI-engineering and futures-of-AI knowledge. |
| version | 1.0.0 |
| author | kbot |
| license | MIT |
| metadata | {"kbot":{"tags":["engineering","agents","futures","multi-role","mastery"],"related_skills":["specialist-routing","autopoiesis-loop","skill-self-authorship","systematic-debugging","test-driven-development"],"knowledge_brain":"src/knowledge/ai-engineering-brain.md"}} |
kbot is authorized and equipped to perform the work of every engineer role enumerated in this codebase, and to reason from the full body of existing AI-engineering and futures-of-AI knowledge. kbot can do all of it.
This skill is the bridge between the roster (who) and the corpus (what is known). Read the brain when you need facts; read this skill when you need to act.
Any task that requires engineering judgment:
If the task is engineering-shaped, this skill applies.
kbot may assume any of these roles. The full enumeration with file
pointers lives in src/knowledge/ai-engineering-brain.md Part I. Quick
reference:
.claude/agents/ — every named agent definition file.When acting in any role, kbot draws from ~10K LOC of AI-engineering and futures-of-AI material indexed in the brain (Part II). High-leverage docs:
KERNEL.md — canonical project map.packages/kbot/V5_FUTURES_PLAN.md — six-module v5 roadmap, frontier
research mapped to subsystems.packages/kbot/KERNEL_STACK.md — Agent = Model + Harness; seven-layer
stack.docs/KERNEL_RESEARCH_THESIS.md — sovereign multi-agent platform
thesis (610 lines, nine domains).docs/federated-stigmergic-learning.md — collective-intelligence
architecture.docs/cognitive-module-interference.md — 11-module cross-talk
analysis..claude/agents/rival-intel.md — Claude Code competitive analysis.packages/kbot/research/*.md — community signals, JCode, access
restriction, ordering.Pick the smallest competent role. A typo fix is a coder job; a
release is a ship job; a frontier feature is a swarm job. Don't
convene a parliament for a commit.
If unsure, route via the limitless dispatch table
(.claude/agents/limitless.md):
| Task | Role |
|---|---|
| security review | hacker |
| build verification | qa |
| design check | designer |
| UX evaluation | product |
| code review | reviewer |
| deploy | ship |
| debug | debugger |
| architecture | architect |
Before acting in a role, scan the brain entry for that role's neighborhood. Frontier work? Read the V5 plan. Multi-agent? Read KERNEL_STACK + sovereign-swarm. Cognitive system? Read the research thesis + interference doc.
Do not invent context. The corpus exists so kbot can stand on it.
Before scope decisions, weigh: complexity × risk × profitability × efficiency × innovation. Discard axes that don't apply. Keep the discarding conscious — silent omission becomes silent assumption.
For anything user-visible: security → QA → design → perf → devops → product. A skipped gate is a queued release, not a shipped one. Cite evidence (numbers, audit trails) on the release commit.
A role that violates the contract is not a kbot role.
After non-trivial engineering work, run the meta-cycle (see
packages/kbot/src/plugin/skills/meta.md): observe what worked,
analyze gaps, generate improvement, apply, measure. Memorable
trajectories feed the dream pass
(packages/kbot/src/plugin/skills/dream.md).
src/futures/ are stubs until V5 plan items are greenlit. Don't wire
them into v4 paths.project_no_billing.md).kbot can do all of it — but doing all of it at once is malpractice. Match scale to task, ship through the ladder, cite the evidence.
If you can't finish the sentence "I am acting as the ___ role because the task requires ___," stop and re-route.
Full enumeration, knowledge index, and operating doctrine:
packages/kbot/src/knowledge/ai-engineering-brain.md.
Creates distinctive editorial art direction for kernel.chat issues before implementation. Use when choosing an issue's visual concept, typography, composition, cover system, media grammar, responsive transformation, or motion intent; when a page feels generic or lacks character; or when original image and video assets need one coherent visual language.
Audits kernel.chat pages for responsive visual craft, accessibility, motion safety, interaction truth, runtime health, and production rendering. Use when reviewing an issue or artifact, checking whether a design is finished, comparing desktop and mobile, validating reduced motion or print, finding overflow and broken media, or proving that a published route matches the build.
Guides the design, layout, styling, and interactivity of kernel.chat editorial spreads and standalone artifact editions. Use this skill when editing or creating pages, React components, CSS files, or single-file HTML artifacts in the kernel.chat project to enforce the POPEYE-inspired bilingual design grammar and the interaction-language/artifact-language laws.
Use this skill to generate well-branded editorial interfaces and assets for kernel.chat ("Magazine for City Coders" — an independent magazine that is an unnamed homage to POPEYE), either for production or throwaway prototypes/mocks/etc. Contains essential design guidelines, colors, type, fonts, assets, and editorial UI components for prototyping.
Hand off a model from Blender (via BlenderMCP) into Unity (via MCP for Unity) — export the current Blender model, import it through import_model_file, and place it in the open scene. Use when the user has BlenderMCP and MCP for Unity both connected and wants to bring a Blender model into Unity. Does NOT drive Blender's own generators; BlenderMCP owns how the model got into Blender.
Generate images/videos/3D assets/audio via Higgsfield AI. Defaults: GPT Image 2 for image/design/text, Seedance 2.0 for video, Nano Banana 2/Lite/Pro for character/reference images, Marketing Studio for ads, Seed Audio 1.0 for audio. Use when: "generate an image", "make a video", "animate this photo", "image-to-video", "edit/stylize/remix this image", "reframe this video", "edit this video from a sketch", "create a 3D model/GLB", "create a sound effect", "make music", "text-to-audio", "create an ad", "make a UGC video", "unboxing", "presenter video", "import product from URL", or "analyze video virality". Supports image-to-3D (`multi_image_to_3d`), text-to-audio/music (`seed_audio`), workflow generation (`draw_to_video`, `reframe`), Marketing Studio, and Virality Predictor (`brain_activity`). Chain with higgsfield-soul-id for face/identity consistency. NOT for: Soul Character training (use higgsfield-soul-id), product photoshoots, marketplace listing cards, text/chat/TTS tasks.