ruvnet/metaharness
🛠️ The meta-harness for AI agents — scaffold your own focused, branded agent harness with its own npx CLI, MCP server, memory, learning loop, and witness-signed releases. Works with Claude Code, Codex, pi.dev, Hermes, OpenClaw, and RVM (har
ARCHETYPE
Skill Collector
Ten-plus skills loaded on demand. A procedural-knowledge library.
Copy this rig
# review before running: this installs third-party code $ claude mcp add kimi-k3-harness -- npx -y kimi-k3-harness@latest mcp start $ claude mcp add code_index -- npx -y kimi-k3-harness@latest mcp index $ npx degit ruvnet/metaharness/kimi-k3-harness/.claude ./rig-metaharness # inspect, then merge into .claude/
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add kimi-k3-harness -- npx -y kimi-k3-harness@latest mcp start $ claude mcp add code_index -- npx -y kimi-k3-harness@latest mcp index
$ npx degit ruvnet/metaharness/.claude-plugin/skills/compare-harnesses .claude/skills/compare-harnesses $ npx degit ruvnet/metaharness/.claude-plugin/skills/create-harness .claude/skills/create-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/diag-harness .claude/skills/diag-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/example-harness .claude/skills/example-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/harness-secrets .claude/skills/harness-secrets $ npx degit ruvnet/metaharness/.claude-plugin/skills/list-templates .claude/skills/list-templates $ npx degit ruvnet/metaharness/.claude-plugin/skills/oia-manifest .claude/skills/oia-manifest $ npx degit ruvnet/metaharness/.claude-plugin/skills/publish-harness .claude/skills/publish-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/repo-genome .claude/skills/repo-genome $ npx degit ruvnet/metaharness/.claude-plugin/skills/score-harness .claude/skills/score-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/threat-model .claude/skills/threat-model $ npx degit ruvnet/metaharness/.claude-plugin/skills/upgrade-harness .claude/skills/upgrade-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/validate-harness .claude/skills/validate-harness $ npx degit ruvnet/metaharness/.claude-plugin/skills/verify-witness .claude/skills/verify-witness $ npx degit ruvnet/metaharness/kimi-k3-harness/.claude/skills/evolve .claude/skills/evolve $ npx degit ruvnet/metaharness/kimi-k3-harness/.claude/skills/plan-change .claude/skills/plan-change
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Bash(git push*)",
"Bash(rm -rf*)"
]
}
} MCP servers (2)
| server | source | est. tokens |
|---|---|---|
| kimi-k3-harness | npm | 2.5k |
| kimi-k3-harness · "code_index" | npm | 2.5k |
Skills (16)
compare-harnessescreate-harnessdiag-harnessexample-harnessharness-secretslist-templatesoia-manifestpublish-harnessrepo-genomescore-harnessthreat-modelupgrade-harnessvalidate-harnessverify-witnessevolveplan-change
Slash commands (2)
/doctor/review-diff
Permissions
deny (4)
Read(./.env)
Read(./.env.*)
Bash(git push*)
Bash(rm -rf*)
ask (0)
—
allow (8)
Bash(npx kimi-k3-harness*)
mcp__kimi-k3-harness__*
mcp__code_index__*
Bash(npm test*)
Bash(npm run*)
Bash(git diff*)
Bash(git status*)
Bash(git log*)
Similar rigs
r1ckyIn/harness-evolve
Self-iteration engine for Claude Code — automatically detects usage patterns and routes optimization recommendations to hooks, skills, rules, CLAUDE.md, memory, and settings
Minimalist 2.3k tok ·
Cherwayway/agent-context-patch
Turn recurring Claude Code and OpenAI Codex mistakes into durable, reviewable workspace memory with safe local context patches.
Minimalist 562 tok ·
manpreet171/slopguard
Security guardrails for vibe coding. Blocks insecure AI-generated code at the moment it's written - hardcoded secrets, SQL injection, hallucinated packages, and agent supply-chain attacks.
Minimalist 2.4k tok ·
alliecatowo/claude-code-darwin
a self learning harness plugin for claude, inspired by mimo code
Pragmatist 428 tok ·