Skills, prompts, AGENTS.md files and MCP servers that Codex reads. Skills are shared with Claude Code; the rest is Codex-native. 282,383 indexed, measured the same four ways as everything else here.
A bilingual guide for planning research and exploring important scientific questions through Socratic questioning, where questions help clarify ideas and assumptions. It follows the idea that research is a conversation with existing knowledge and other researchers.
Use when an architect must run the detailed-design portion of technical design, preserve the full lifecycle-stage doctrine for detailed design, and define the concrete runtime, project, flow, data, interface, non-functional, security, recovery, hardware, and implementation-plan surfaces needed for execution.
A structured workflow for software engineering tasks. It routes work through one of three levels, from a quick path to a full 11-step process, with checklists, approval gates, rollback rules, and records of decisions.
A general reverse-engineering toolkit for Windows programs, Android apps, iOS apps, web interfaces, and APIs. Reverse engineering means examining software to understand how it works and, where appropriate, how to modify and rebuild it.
Quick-reference card for all ponytail modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /ponytail-help, "ponytail help", "what ponytail commands", "how do I use ponytail".
Lets the agent drive a real browser: open pages, click, type, take screenshots and read the accessibility tree, using Playwright. Runs locally from the @playwright/mcp npm package.
Use when modifying the Light Dashboard custom/ variant. Covers the custom mini-CMS engine, layout and color customization, adding/removing pages, HTMX navigation, and CMS-level SMQ page scopes.
Set up and configure Context Harness standalone local agents, declarative tool bindings, and reusable agent workflows. Use when creating, validating, running, or debugging project agents; do not use for legacy MCP prompt-only personas unless migrating them.
Move one experiment through RUN → VERIFY → DELIVER with the least control work that safely advances the user's outcome. This is the compute/control layer and owns one compute attempt. When the task spans several runs, nodes, independent side-effect lanes, or paper-wide closeout, route scheduling and state projection…
MCP server "notebooklm-mcp-structured" as configured in paolodalprato/notebooklm-mcp-structured. Runs locally from the notebooklm-mcp-structured npm package.
★not rated 42 2mo agoA
tokens not measured
originalMITarchived
Call local Grok CLI, Kimi Code CLI, and Claude Code CLI with the strongest current default models and native tools. Use this qiaomu skill workflow when the user wants to run Grok 4.5 for X/web search, image/video generation, research, or multi-tool agent work; run Kimi K3 (1M) for frontend UI/CSS/React/Vue work; run…
Deep autoscaling diagnostics and tuning for three scaling paths: 1) Horizontal Pod Autoscaler (HPA), 2) Vertical Pod Autoscaler (VPA), 3) Karpenter node provisioning.
Rope-first workflow for ANY mechanical Python rename/move (functions/methods/classes/variables/modules/packages), updating imports/references via scripts/roperefactor.py; use even for small renames; only fallback if rope fails/dynamic imports.
Guide for adding new wsh commands to Wave Terminal. Use when implementing new CLI commands, adding command-line functionality, or extending the wsh command interface.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: