What is eval-kernel?
eval-kernel is a Claude Code mod by kzarzycki that a persistent JavaScript kernel the model drives with one `eval` tool: state survives between cells, and cells call Claude Code tools, models and subagents. It is listed under Agents & automation and has 1 GitHub stars.
| Type | Claude code mod |
|---|---|
| Author | kzarzycki |
| Repository | kzarzycki/claude-mods |
| Popularity | 1 GitHub stars |
| Category | Agents & automation |
| Security | Review before installing (64/100) |
| Rank | #1,930 of 5,272 Claude Code mods; #146 in Agents & automation |
| Last commit | 2026-10-09 |
| Hooks | 5 events |
| Sees | every prompt, every tool call, mcp__eval-kernel__eval calls |
| Draws on | nothing |
Is eval-kernel safe to install?
eval-kernel scores 64/100 on the MCPMods security check: Review before installing. Things to review: sends data over the network; watches every prompt or event. It passed 3 of 8 checks.
Mod footprints come from claude plugin validate, which lists every event a mod hooks and every API call it makes without running it.
| Check | Result |
|---|---|
| Sends data over the network | Review |
| Starts no processes | Pass |
| Writes no files or config | Pass |
| Doesn't read environment variables | Pass |
| Watches every prompt or event | Review |
| Drives Claude (model calls, prompts) | Note |
| Validator reported problems | Note |
| No license detected | Note |
API calls in its code
| Call | What it means |
|---|---|
$.agent.spawn | drives Claude |
$.clock.sleep | |
$.command.register | |
$.decision.ask | |
$.http.fetch | network requests |
$.model.complete | drives Claude |
$.process.spawn | |
$.session.cwd | |
$.session.id | |
$.tool.call | drives Claude |
$.tool.list | |
$.tool.register | drives Claude |
$.ui.log | draws interface |
How do I install eval-kernel?
- Check your Claude Code version. Mods need Claude Code 2.1.287 or later and are on by default.
claude --version
- Read its footprint. This lists every event it hooks and every API call it makes, without running it.
git clone https://github.com/kzarzycki/claude-mods claude plugin validate ./claude-mods/plugins/eval-kernel
- Install it. Run this in your terminal, then /reload-plugins in an open session.
claude plugin marketplace add kzarzycki/claude-mods && claude plugin install eval-kernel@claude-mods
Or ask Claude to set it up
Paste this into Claude Code and it installs eval-kernel for you:
Install the "eval-kernel" Claude Code mod from https://github.com/kzarzycki/claude-mods/tree/main/plugins/eval-kernel. Run claude plugin validate on it first and show me what it can access before you enable it.
More from kzarzycki/claude-mods
This repository has 8 Claude Code mods on MCPMods.
- agent-hub: Agent hub: a pane listing this session's subagents with status, model, turns and tokens, and a box to steer a running on (Low risk)
- auto-effort: Auto effort: the decision model rates each prompt and the turn it starts runs at that effort. (Low risk)
- compact-methods: oh-my-pi compaction methods: shake (move old tool output to files, no model call) and handoff (continue from a handoff d (Review before installing)
- decision-model: A decision model for other mods ($.decision.ask): typed questions about a state, answered by any System One endpoint (Je (Review before installing)
- resume-on-stop: Unexpected stops: when a turn ends right after saying it will act, the decision model spots it and one turn resumes the (Review before installing)
- self-compact: Compaction at task boundaries: the model can ask to compact once its turn ends. Above a fill threshold it is asked wheth (Review before installing)
- ttsr-rules: oh-my-pi TTSR rules: a regex rule denies the tool call that would break it, with the rule as the reason; a question rule (Review before installing)