~ / rigs / raphaelchristi / harness-evolver

raphaelchristi/harness-evolver

Automated harness evolution for AI agents. A Claude Code plugin that iteratively optimizes system prompts, routing, retrieval, and orchestration code using full-trace counterfactual diagnosis. Based on Meta-Harness (Lee et al., 2026).

↗ GitHub ★ 53 mit updated 6mo ago personal setup Claude Code Claude plugin
share on X
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
CONTEXT TAX · EVERY TURN
~4.0k tokens
Moderate · median rig: 2.3k · breakdown
GUARDRAILS
0/5
No committed guardrails · details

Copy this rig

# review before running: this installs third-party code
$ npx degit raphaelchristi/harness-evolver/.claude ./rig-harness-evolver  # 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 raphaelchristi/harness-evolver/.claude/skills/dev-dry-run .claude/skills/dev-dry-run
$ npx degit raphaelchristi/harness-evolver/.claude/skills/dev-release .claude/skills/dev-release
$ npx degit raphaelchristi/harness-evolver/.claude/skills/dev-validate .claude/skills/dev-validate
$ npx degit raphaelchristi/harness-evolver/skills/certify .claude/skills/certify
$ npx degit raphaelchristi/harness-evolver/skills/deploy .claude/skills/deploy
$ npx degit raphaelchristi/harness-evolver/skills/evolve .claude/skills/evolve
$ npx degit raphaelchristi/harness-evolver/skills/health .claude/skills/health
$ npx degit raphaelchristi/harness-evolver/skills/setup .claude/skills/setup
$ npx degit raphaelchristi/harness-evolver/skills/status .claude/skills/status
$ curl -fsSL --create-dirs -o .claude/agents/harness-architect.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-architect.md
$ curl -fsSL --create-dirs -o .claude/agents/harness-consolidator.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-consolidator.md
$ curl -fsSL --create-dirs -o .claude/agents/harness-critic.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-critic.md
$ curl -fsSL --create-dirs -o .claude/agents/harness-evaluator.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-evaluator.md
$ curl -fsSL --create-dirs -o .claude/agents/harness-proposer.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-proposer.md
$ curl -fsSL --create-dirs -o .claude/agents/harness-testgen.md https://raw.githubusercontent.com/raphaelchristi/harness-evolver/main/agents/harness-testgen.md

This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 7,204 rigs:

{
  "permissions": {
    "deny": [
      "Read(./.env)",
      "Read(**/.env)",
      "Read(~/.ssh/**)",
      "Bash(rm -rf *)",
      "Read(**/*.pem)",
      "Bash(rm -rf /)",
      "Bash(git push --force:*)",
      "Bash(sudo *)",
      "Read(.env)",
      "Bash(rm -rf /*)",
      "Read(./.env.*)",
      "Read(~/.aws/**)",
      "Bash(git push --force*)",
      "Bash(rm -rf:*)",
      "Read(**/*.key)",
      "Read(**/.env.*)",
      "Bash(sudo:*)",
      "Bash(git reset --hard*)",
      "Bash(git reset --hard:*)",
      "Read(.env.*)"
    ],
    "ask": [
      "Bash(git push:*)",
      "Bash(git push *)",
      "Bash(git commit:*)",
      "Bash(rm *)",
      "Bash(rm:*)",
      "Bash(git rebase *)",
      "Bash(wget *)",
      "Bash(npm publish:*)",
      "Bash(git commit *)",
      "Bash(gh pr merge *)"
    ]
  }
}

Skills (9)

Subagents (6)

harness-architect
model: opus
|
harness-consolidator|
harness-critic|
harness-evaluator|
harness-proposer|
harness-testgen|

Similar rigs

copied ✓