Genshijin Skills - Setup Guide¶
Source: interfacex-co-jp/genshijin GitHub: interfacex-co-jp/genshijin Skills: 7 skills · 19.2K total installs Category: Agent Communication First Seen: April 7, 2026 (catalogued August 18, 2026 sweep) Quality Tier: 🟢 Production - Gen Agent Trust Hub Pass, Socket Pass, and Snyk Pass on the genshijin flagship; 310 GitHub stars; #2 on the hot page at sweep time
Genshijin (原始人, "primitive man") is a Japanese concise-reply framework for coding agents. The core skill enforces brutal brevity - all technical content kept, all filler removed - with three persistent modes (通常 normal, 丁寧 polite, 極限 extreme) switched via a slash command, and it stays active across turns, resisting politeness regression and filler drift. Six companion skills extend it into a full workflow: genshijin-commit (concise conventional commits), genshijin-review (terse code review), genshijin-compress (context compression), genshijin-help, genshijin-stats, and genshijin-crew (multi-agent crew coordination). The cluster sits at #2 on the skills.sh hot page with genshijin-commit at +13 installs in one hour.
Installation¶
npx skills add interfacex-co-jp/genshijin
Individual skills install with the explicit repo form:
npx skills add https://github.com/interfacex-co-jp/genshijin --skill genshijin
Prerequisites¶
| Requirement | Details |
|---|---|
| Node.js + npx | For the skill installer |
| Japanese-capable agent | The SKILL.md content ships in Japanese; the framework works for any language output but the instruction text is Japanese |
What It Provides¶
| Skill | Installs | Purpose |
|---|---|---|
| genshijin | 3.0K | Core concise-reply framework with persistent mode switching (通常/丁寧/極限) |
| genshijin-commit | 2.7K | Concise conventional commit messages |
| genshijin-compress | 2.7K | Context compression for long sessions |
| genshijin-review | 2.7K | Terse code review feedback |
| genshijin-help | 2.7K | Usage help for the framework |
| genshijin-stats | 2.7K | Reply-length and mode statistics |
| genshijin-crew | 2.7K | Crew coordination across multiple agents |
Quick Start¶
- Install:
npx skills add interfacex-co-jp/genshijin - The core skill activates automatically; switch modes with
/genshijin 丁寧|通常|極限 - Disable with 「原始人やめて」or 「通常モード」
CorpusIQ Use Cases¶
| Use Case | How |
|---|---|
| Token discipline | genshijin and genshijin-compress align with our field-filtering and memory-compaction token strategy |
| Consistent commit style | genshijin-commit enforces concise conventional commits across agent-built repos |
| Multi-agent coordination | genshijin-crew maps to multi-agent orchestration workflows |
| Reply hygiene | Mode persistence prevents filler drift on long-running agent sessions |
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations / Verification¶
- Security audits on the genshijin flagship: Gen Agent Trust Hub Pass, Socket Pass, Snyk Pass
- Publisher-page total verified (19.2K across 7 skills); 310 GitHub stars as of the sweep
- Below the 20K install guide bar - drafted on #2 hot-page position, +13 one-hour momentum on genshijin-commit, and Japanese corporate publisher (interfacex-co-jp)
- SKILL.md instruction text is Japanese - the skills.sh excerpt and install content carry over untranslated and agents must handle the original-language instructions (same handling as the wecomteam suites)
npx skills add interfacex-co-jp/genshijin # verify install works
Related¶
Powered by CorpusIQ