TeFuirnever
Useromm (oh-my-matrix)：OpenClaw-native orchestration extension suite.
Categories
Indexed Skills (16)
dynamic-workflows
Multi-agent workflow orchestration for tasks that exceed a single agent — use proactively for parallel, multi-perspective, or large-scope work in ANY domain (development, review, task processing): fan out, then _refute_ to keep only what survives. Use when a task needs 3+ independent perspectives, 10+ files to audit/migrate/review/refactor, or a productive task with a natural refute gate (implement-then-review, generate-then-filter, select-via-judge). Also trigger on: parallel agents, multi-file changes, cross-module refactor, bulk migration, deep audit across many files, independent subtasks, 并行 agent, 多视角, 审计, 迁移, 重构, 多文件. Falls back to bounded direct-session plans for small tasks.
arch-doc
End-to-end authoring and calibration of an architecture design document from upstream sources. MUST trigger for 写架构设计文档, 架构说明书, ARCH 文档, 软件实现架构说明书, 架构文档补时序图, PlantUML 时序图（架构文档内）, 交叉校准, 回源核对, 上游研读, 上游真源, 最佳实践评审, 架构评审对标, 架构文档模板, PRD 模板, 产品需求文档模板, 功能详细设计模板, 校准架构文档, 质量场景卡, or whenever a repository needs a software-implementation-architecture document drafted, calibrated against upstream design docs, or reviewed against industry best practices. Distilled from the MCC ARCH-2026-001 authoring pipeline.
autoresearch
Wrap Karpathy's autoresearch environment so an agent can run, resume, and report automatic nanochat training experiments inside a user-owned pinned checkout. Requires an NVIDIA GPU and uv; vendors no upstream bytes. Use when the user asks to run or check autoresearch experiments.
azhou-setup
Plan and explicitly apply Azhou AI Hub skill installation or receipt-owned repair, migration, and uninstall. Use for checkout-assisted setup and managed lifecycle operations with an exact target root.
azhou-verify
Run and report the authoritative Azhou AI Hub repository verification gate. Use before completion, handoff, commit, pull request, release, or when full-codebase evidence is requested.
eli5
Explain a topic like I'm 5 with one HTML artifact of big pictures and few words. Use when the user types /eli5 <topic> or asks for a dead-simple picture explainer of how something works.
excalidraw-diagram
Build or edit accurate, editable Excalidraw scenes; render the real scene, inspect the image, run deterministic layout/style checks, and deliver source plus requested exports. Use for workflows, architectures, sequences, data flows, concept maps, or existing .excalidraw files. Supports native JSON, offline Mermaid/SVG conversion, official component libraries, CJK-safe SVG/PNG export, and optional interactive preview.
session-insights
Analyze local agent session stores into fact-bound usage insight reports — session counts, active days, hour/project distribution, tool ranking, friction signals. Use when the user asks how they actually use their coding agent, wants a weekly report, or asks to be roasted on real numbers; read-only, aggregates only, raw transcripts never leave the machine.
super-lavish
Renamed from lavish (the old name still triggers this skill). Turn complex or visual agent responses into rich, reviewable HTML artifacts that users can annotate and send feedback on through the Lavish Editor CLI, and relay a PRD, RFC, design spec, or technical plan with comments, selected-text annotations, feedback disposition, and next-owner state inside one portable HTML file. Use for visual artifacts, HTML explainers, interactive prototypes, review surfaces, product or technical plans, team spec review or transfer, comparisons, diagrams, tables, code views, reports, slides, or browser-based feedback loops.
super-llm-wiki
Renamed from llm-wiki (the old name still triggers this skill). Build, query, lint, migrate, and maintain a private project Markdown wiki when verified architecture, decisions, debugging facts, or conventions must persist across sessions. Do not use it for global memory, ephemeral scratch notes, secrets, or unreviewed transcripts.
super-repo-pedant
Renamed from repo-pedant (the old name still triggers this skill). Reconcile repository knowledge at explicit task close. MUST trigger for sync up, tidy or clean up docs, update memory, /sync, /neat, /repo-pedant, 同步一下, 整理文档, 整理一下, 更新记忆, 梳理一下, 收尾, 这个阶段做完了, 新人能直接上手, stale docs, conflicting memories, clean handoff, or bare tidy/整理 in development context, 检查项目技术债务, 仓库健康检查, repo health check. An audit phrasing invites a repository-level review; a request scoped to a single file or task stays ordinary work. Inferred completion only reminds; ordinary implementation that merely mentions or edits this skill does not authorize closeout.
harness
How to use and extend the oh-my-matrix development harness. Run the gates locally before pushing, locate each config, add a new gate, and follow the commit convention. Use when onboarding to OMM, running checks before a push, or adding/modifying a quality gate. Triggers: 'run the harness', 'pnpm check', 'pnpm verify', 'add a lint rule', 'add a gate', 'how do gates work', 'commit format', 'harness'.
work-with-pr
Full PR lifecycle for oh-my-matrix in a task-owned git worktree — implement with evidence-bound QA, open a reviewer-readable PR against `master`, then run an unbounded verify loop over three gates (CI / review-work / docs-build) until all pass, then merge by default and clean up the worktree. Decomposes work into the smallest atomic, independently-mergeable PRs and builds the independent ones concurrently. Use whenever implementation work needs to land as a PR. Triggers: 'create a PR', 'implement and PR', 'land this', 'work-with-pr', 'implement issue', 'split into atomic PRs', 'parallel PRs', even 'implement X' when the context implies PR delivery.
azhou-doctor
Diagnose Azhou AI Hub checkout, package, explicit install-target, or Treehouse lease health without mutation. Use for health checks, broken installs, environment diagnostics, or support verification.
azhou-info
Report provable Azhou AI Hub checkout information or revision facts. Use for project info, installable repository inventory, support facts, version, commit, branch, or dirty-state questions.
super-caveman
Enhanced Caveman skill that combines the original terse-mode core, six companion routes, and the complete pinned i-have-adhd output-behavior contract in one installable package. Use for "super caveman", "caveman mode", "be brief", token efficiency, /super-caveman, /caveman, /cavecrew, /caveman-commit, /caveman-review, /caveman-compress, /caveman-help, /caveman-stats, commit-message requests, concise PR review, compact delegation, prose compression, or exact usage statistics.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.