chris-code¶
chris-code turns Claude Code into an opinionated software-engineering workflow rather than a free-form chat assistant. It routes any non-trivial change through a fixed pipeline — brainstorm intent, write a lean spec, hand off a thin plan, dispatch a coder subagent per task, run staged review, and gate drift before it reaches main. It ships 25 skills, 14 dedicated agents, and review gates that re-read the actual code instead of trusting an agent's summary.
When to use it¶
- You want Claude to design before it codes — settle what and why before how.
- You want the main context window kept clean by offloading focused work to subagents.
- You want agent-generated drift caught by review gates before it lands.
- You work primarily in Python or Rust (the coder and review agents are language-scoped), though the workflow skills are language-agnostic.
If you just want a quick one-off answer, you don't need chris-code. It earns its keep on changes substantial enough to deserve a design and a review.
Install¶
chris-code is distributed through a personal Claude Code marketplace. Add it to ~/.claude/settings.json:
{
"extraKnownMarketplaces": {
"chris-code": {
"source": { "source": "github", "repo": "chris-santiago/claude-plugins" }
}
},
"enabledPlugins": {
"chris-code@chris-code": true
}
}
Claude Code fetches and caches the plugin automatically. (The repo also ships ml-lab and ml-journal; enable those separately if you want them.) Full install options — local directory, manual cache — are in the repository README.
Where to go next¶
- Start with the tutorial → — take one change from idea to merged, end to end.
- Browse the how-to guides → — task recipes: fix a bug, run a determined change, review a branch, run a quality campaign.
- Read the reference → — the full catalog of skills and agents, and how dispatch works.
- Understand the design → — why a fixed pipeline, what the gates actually prove, and when to dispatch vs. stay in-session.