~ / rigs / engn-dev / meta-harness-studio

engn-dev/meta-harness-studio

Author one AI coding harness once, project it onto Claude Code, Codex, OpenCode, Cline, Kilo Code & Pi — then let it improve itself from execution traces (arXiv:2603.28052).

↗ GitHub ★ 1 MIT updated 3mo ago personal setup Claude CodeCodex
share on X
ARCHETYPE
Minimalist
Lean instructions, few tools, tiny context tax. Lets the model think.
CONTEXT TAX · EVERY TURN
~1.4k tokens
Featherweight · median rig: 2.3k · breakdown
GUARDRAILS
3/5
Protects secrets · No YOLO mode · Sandbox or ask-first rules · details

Copy this rig

# review before running: this installs third-party code
$ npx degit engn-dev/meta-harness-studio/.claude ./rig-meta-harness-studio  # 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.

$ npx degit engn-dev/meta-harness-studio/.agents/skills/example-skill .claude/skills/example-skill
$ npx degit engn-dev/meta-harness-studio/examples/demo-project/.harness/skills/api-conventions .claude/skills/api-conventions

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*)",
      "Edit(./dist/**)",
      "Write(./dist/**)"
    ]
  }
}

Skills (2)

Slash commands (1)

/review

Permissions

deny (3)
Read(./.env*)
Edit(./dist/**)
Write(./dist/**)
ask (0)
—
allow (0)
—

Similar rigs

copied ✓