本文へ移動
cccskills
無料GitHub で公開

dsh-setup

Install DSH on a fresh Mac. Runs sovereign-setup-mac.sh, renders .env from secrets backend, applies full sovereign lockdown, probes machine, and pulls right-sized Ollama models — one guided flow that leaves the node hardened and ready for Phase-2 spore.sh activation.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md5.2 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

You are bootstrapping a Dome Sovereign Hub (DSH) node on a fresh Apple Silicon Mac. The user has already cloned garochee33/DSH to $HOME/DSH (or set $DOME_ROOT). Your job is to run the end-to-end setup without leaving anything halfway.

1. Preflight

Confirm before doing anything destructive:

  • $DOME_ROOT resolves to a real directory (default: $HOME/DSH). If not, stop and ask.
  • macOS version is ≥ 13 (sw_vers -productVersion). Windows users must run scripts/sovereign-setup-windows.ps1 instead — bail out.
  • xcode-select -p returns a path. If not, run xcode-select --install and exit — tell the user to re-run after it completes.
  • sudo -v prompt works (fail early if the user can't authorize).

2. Run the base setup

cd "$DOME_ROOT"
bash scripts/sovereign-setup-mac.sh

This is the 20-step Homebrew/pyenv/nvm/VSCode/DB bootstrap. It is idempotent — safe to re-run. Near the end it prompts for an AI-assistant pick; choose based on user preference or skip.

When it finishes, verify:

  • command -v brew && command -v pnpm && command -v pyenv all print paths.
  • "$DOME_ROOT/.venv/bin/python" --version reports the version pinned in $DOME_ROOT/.python-version (currently Python 3.14.x).
  • cat "$DOME_ROOT/.env" exists (copied from .env.example by the setup script if it wasn't already there).

3. Render secrets into .env (if template backend is in use)

If .env.template has {{keychain:...}} or {{pass:...}} markers:

bash "$DOME_ROOT/scripts/render-env.sh"

This replaces markers with values from macOS Keychain (preferred) or pass (GPG). Confirm with head -5 "$DOME_ROOT/.env" — there should be no {{...}} tokens left. If any marker remains, list them for the user to provision before continuing.

4. Apply sovereign lockdown

bash "$DOME_ROOT/scripts/lock-down.sh"           # kill phone-home daemons, pin DNS
bash "$DOME_ROOT/scripts/lock-down-phase2.sh"    # firewall + stealth
bash "$DOME_ROOT/scripts/lock-down-phase3.sh"    # pf anchor for telemetry endpoints
bash "$DOME_ROOT/scripts/lock-down-phase4.sh"    # daemon audit + permanent removal
bash "$DOME_ROOT/scripts/pf-reload.sh"           # reload pf with the anchor

Each script will prompt for sudo once and keep the cached credential alive while it runs. If any exits non-zero, STOP — do not proceed to steps 5+ until the lockdown is clean.

5. Probe the machine

"$DOME_ROOT/.venv/bin/python" "$DOME_ROOT/scripts/machine-probe.py"

Write-out lands at agents/core/.mesh/machine.json. Report the summary line to the user:

Apple M5 Pro · 18 cores (6S + 12P) · 48.0 GB RAM · 40+ TOPS NPU · tier=sovereign

6. Pull right-sized local LLM weights

bash "$DOME_ROOT/scripts/ollama-init.sh"

This reads the tier detected in step 5 and pulls the matching Ollama models. On a sovereign tier (≥18 GB RAM) expect qwen2.5-coder:14b + llama3.1:8b + nomic-embed-text. Confirm with ollama list.

7. Initialize state + ingest KB

cd "$DOME_ROOT"
source .venv/bin/activate
python3 scripts/register-claude.py    # creates dome.db tables: agents, skills, tools, stack
python3 scripts/ingest.py             # indexes kb/, logs/, docs into ChromaDB

After this: sqlite3 db/dome.db "SELECT COUNT(*) FROM skills" should return ≥ 8. python3 -c "import chromadb; print(chromadb.PersistentClient('db/chroma').get_collection('dome-kb').count())" should return ≥ 1500.

8. Verify all protocols

python3 scripts/pre-spore-verify.py   # expect: 27/27 READY FOR SPORE.SH
bash scripts/audit.sh                 # expect: all green (or documented yellows)
bash scripts/dome-check.sh            # expect: 0 failures on a fresh install

If any of the three fails, surface the specific failure and do NOT declare setup complete. Common causes:

  • akashic:write+query fails → stale akashic ChromaDB collection. Fix: delete via python3 -c "import chromadb; chromadb.PersistentClient('db/chroma').delete_collection('akashic')" then re-run.
  • dome-check fails on SQLite DB missing → skipped step 7. Go back.
  • audit.sh fails on FileVault OFF → user must enable in System Settings → Privacy & Security. This is not auto-fixable.

9. Report

Give the user a single-block status:

✅ DSH installed at $DOME_ROOT
✅ Lockdown active — firewall + pf anchor + DNS pinned
✅ Machine: <summary_one_liner output>
✅ KB indexed: <N> chunks in dome-kb
✅ pre-spore-verify: 27/27
Next: run `spore.sh` with a Trinity SPORE_TOKEN to activate Phase 2 (mesh).

Non-negotiables

  • Never proceed past a failing step.
  • Never commit .env or log output from lockdown scripts.
  • Never run lockdown steps without explicit user sudo approval (the scripts handle it, but if they error-out silently, STOP).
  • If the user has a non-M-series Mac (Intel), warn them — DSH is tested on Apple Silicon; some MPS-specific PyTorch code will not use acceleration.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

ai-sdk

無料

Vercel AI SDK expert guidance. Use when building AI-powered features — chat interfaces, text generation, structured output, tool calling, agents, MCP integration, streaming, embeddings, reranking, image generation, or working with any LLM provider.

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

Algorithm design and analysis skill for DSH; use when working on optimization, data structures, or computational procedures.

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

Design and implement scalable API gateways, RESTful APIs, GraphQL endpoints, WebSocket handlers, and microservice communication patterns. Use when creating API routes, designing endpoints, implementing middleware, or setting up service mesh architectures.

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

best-of-n

無料

Implement a task N ways in parallel and pick the best. Spawns multiple subagents in isolated worktrees, evaluates all candidates, and applies the winner. Use when asked to "best of n", "try multiple approaches", "parallel implementations", "/best-of-n", or "/bon".

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

check

無料

Check your work with a verification subagent. Spawns a verifier that reviews diffs, runs builds and tests, and evaluates correctness. Use when asked to "check work", "verify changes", "self-verify", "/check", "/verify", "/check-work", or "/self-verify".

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

CI/CD pipeline design, GitHub Actions workflows, deployment automation, infrastructure as code, and release management. Use when setting up CI/CD, automating deployments, configuring build pipelines, or managing releases.

日本語の概要は準備中です。原文の説明を表示しています。

garochee33/DSH42026年10月9日 更新

garochee33 のスキルをすべて見る

このスキルの問題を報告する