Ancienttwo/repo-harness
File-backed workflow harness for reliable Claude Code and Codex sessions.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit Ancienttwo/repo-harness/agents ./rig-repo-harness/agents 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 Ancienttwo/repo-harness/assets/skills/obsidian-memory .claude/skills/obsidian-memory $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-archimap .claude/skills/repo-harness-archimap $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-cross-review .claude/skills/repo-harness-cross-review $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-herdr-portal .claude/skills/repo-harness-herdr-portal $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-issue-flow .claude/skills/repo-harness-issue-flow $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-product .claude/skills/repo-harness-product $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-setup .claude/skills/repo-harness-setup $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-strategy .claude/skills/repo-harness-strategy $ npx degit Ancienttwo/repo-harness/assets/skills/repo-harness-test .claude/skills/repo-harness-test $ npx degit Ancienttwo/repo-harness/assets/skills/windows-python-first .claude/skills/windows-python-first
$ curl -fsSL --create-dirs -o .claude/agents/deep-reasoner.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/deep-reasoner.md $ curl -fsSL --create-dirs -o .claude/agents/deep-worker.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/deep-worker.md $ curl -fsSL --create-dirs -o .claude/agents/explorer.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/explorer.md $ curl -fsSL --create-dirs -o .claude/agents/fast-worker.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/fast-worker.md $ curl -fsSL --create-dirs -o .claude/agents/gatekeeper.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/gatekeeper.md $ curl -fsSL --create-dirs -o .claude/agents/harness-evaluator.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/harness-evaluator.md $ curl -fsSL --create-dirs -o .claude/agents/known-gaps.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/known-gaps.md $ curl -fsSL --create-dirs -o .claude/agents/root-cause-prover.md https://raw.githubusercontent.com/Ancienttwo/repo-harness/main/agents/fleet/root-cause-prover.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Skills (10)
obsidian-memoryrepo-harness-archimaprepo-harness-cross-reviewrepo-harness-herdr-portalrepo-harness-issue-flowrepo-harness-productrepo-harness-setuprepo-harness-strategyrepo-harness-testwindows-python-first
Subagents (8)
| deep-reasoner model: opus | Architecture research and judgment executor on Opus at xhigh effort. Use for architecture proposals, complex multi-step reasoning, and high-risk analysis — design trade-offs, risky migrations, securit |
| deep-worker model: opus | Heavy execution worker on Opus at high effort. Use for hard, well-scoped execution — cross-module refactors, tricky concurrency or state fixes, changes that must land right in one pass; may run scoped |
| explorer model: sonnet | Read-only codebase explorer on Sonnet at medium effort. Use to locate files, symbols, call paths, tests, uncertainties, and candidate affected surfaces before implementation or judgment. Returns `FIND |
| fast-worker model: sonnet | Fast execution worker on Sonnet at high effort. Use for well-scoped implementation, tests, refactoring, documentation, and mechanical changes; may run scoped typecheck and affected tests in its own wo |
| gatekeeper model: opus | Read-only acceptance reviewer on Opus at high effort. Uses recorded checks for large changes, security/permissions or unresolved uncertainty. Returns a verdict and recommendation. Never edits or publi |
| harness-evaluator model: opus | Disposable-state harness behavior evaluator on Opus at high effort. Use to run the existing skill-evaluation or adoption-inspection profiles inside a complete disposable repo and HOME, then report reg |
| known-gaps | — |
| root-cause-prover model: opus | Bugfix diagnosis specialist on Opus at xhigh effort. Proves a cause with a real reproduction and command evidence. Returns CONFIRMED, LIKELY or BLOCKED. Never fixes production source or changes gates. |
Similar rigs
aimpact-inc/repo-harness
File-backed workflow harness for reliable Claude Code and Codex sessions.
Orchestrator 6.6k tok · · Q10
jeroromano/claude-code-orchestration-template
—
Pragmatist 2.4k tok · · Q8
thomascasali/claude-kb-workflow
Knowledge that crosses your projects: Claude Code native memory is per-project and per-machine, this adds the layer above. Sessions become a wiki; a page reaches the stable cross-project KB only after a numeric bar: 2+ projects, 2+ sessions
Orchestrator 1.1k tok · · Q0
bi-boo/claude-code-config
我的 Claude Code 个人配置——CLAUDE.md 规则、Agent、记忆机制,以及 Ghostty 终端配置
YOLO Cowboy 1.4k tok · · Q16