FrogAi

Xenopus — AI skill for Claude Code

AI community

Specimens from my LLM lab: agents, hooks, instructions, MCP servers, prompts, and skills for Claude, Codex, Gemini, and friends.

How to install Xenopus

This entry records only its repository, not the path inside it, so there is no exact command to give. Open FrogAi/Xenopus and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Xenopus does

Specimens from my LLM lab: agents, hooks, instructions, MCP servers, prompts, and skills for Claude, Codex, Gemini, and friends.

Alternatives in AI

System Prompts Leaks — Extracted system prompts from ChatGPT (GPT-5.4, GPT-5.3, Codex), Claude (Opus 4.6, Sonnet 4.6, Claude Code), G 38.6k ★
TokenTracker — Local-first AI token usage & cost tracker for 31 coding tools incl 1.4k ★
HarnessKit — More than a skill manager — manage skills, MCP servers, plugins, hooks, CLIs, configs, memory & rules across e 416 ★

README

Xenopus

Xenopus, just like the lab frog of the same name, is my experimental Claude + Codex workflow library for high-rigor engineering work. It collects the instructions, specialist agents, reusable skills, and portable config examples behind a deliberately heavyweight engineering posture: maximum quality and performance, with no concern for speed, latency, token cost, or compute cost.

This library is for teams and individuals who want principal-level, FAANG-scale review pressure in their local assistant workflows. It is slightly opinionated: inspect real sources, prefer native Claude/Codex surfaces, validate hard, and keep only durable workflow artifacts that earn their maintenance cost.

At a Glance

Surface	Claude target: `Opus 4.7`	Codex target: `GPT 5.5`
Canonical workflows	75	75
Custom agents	33	33
Skills	46	46

Repository Layout

LICENSE
README.md
Claude/
  Instructions/
    Opus 4.7/
      maximum-quality.md
      settings.example.json
  Agents/
    Opus 4.7/
      .md
  Skills/
    Opus 4.7/
      /
        SKILL.md
Codex/
  Instructions/
    GPT 5.5/
      AGENTS.md
      config.example.toml
      rules/
        default.rules
  Agents/
    GPT 5.5/
      .toml
  Skills/
    GPT 5.5/
      /
        SKILL.md
        agents/
          openai.yaml

How the Pieces Fit

Agents define delegated specialist roles. They are narrow, bounded lanes for implementation, review, research, and validation work where isolated context improves quality.

Config examples show the portable setup shape. They are intentionally examples, not drop-in credentials or machine-specific state. Review paths, permissions, plugins, MCP servers, connector settings, and environment variables before adapting them.

Instructions define the default behavior for a runtime. `maximum-quality.md` is the Claude output style for this stack, and `AGENTS.md` is the Codex-native instruction surfac