Getting Started
Get SigmaShake governance running in your project in under 2 minutes.
SigmaShake Desktop is a tray-icon app for Windows, macOS, and Linux that sets everything up for you with no terminal required — download, double-click, done. Install SigmaShake Desktop →
Download SigmaShake Desktop
The fastest no-terminal path. SigmaShake Desktop is a tray-icon app that runs the governance daemon and opens the dashboard for you.
- All platforms (Windows, macOS, Linux) + direct downloads: sigmashake.com/desktop — Windows installer, macOS
.dmg, Linux.tar.gz(+ AppImage fallback), and the one-linecurlinstall. Free to download; Pro is $5/month, subscribe at sigmashake.com/pricing. - Mac App Store: SigmaShake Desktop on the Mac App Store — free app, sandboxed, auto-updates, macOS 14+ on Apple Silicon. Pro is a $5/month in-app purchase through Apple; per Apple's App Review policy, an existing sigmashake.com subscription does not carry over to this build, so buy Pro here (not on the web) if you're using the Mac App Store version.
- Microsoft Store: SigmaShake Desktop on the Microsoft Store — sandboxed, auto-updates.
Option A — AI Agent Setup (Recommended)
If you use Claude Code, Cursor, Antigravity, or any other MCP-compatible AI coding agent, this is the fastest path. Two commands, then just talk to your agent.
Step 1 — Install the CLI and subscribe
The CLI is free to download and install. A Pro or Enterprise subscription is required to use governed value surfaces. Subscribe at sigmashake.com/pricing, then log in:
ssg auth login
Use the install channel that fits your environment:
- macOS install guide — curl one-liner
- Linux install guide — curl one-liner or Docker/GHCR
- Windows install guide — PowerShell installer
# macOS / Linux
curl -fsSL https://install.sigmashake.com | sh
# Docker / GHCR
docker run --rm ghcr.io/sigmashakeinc/ssg:1.0.7 --version
Step 2 — Connect your AI agent
# Auto-detects your AI client (Claude Code, Antigravity, Cursor, etc.):
ssg init --client=claude-code
# Or specify explicitly:
ssg init --client=antigravity
ssg init --client=gemini
ssg init --client=codex
# Want to see what `ssg init` will change before it runs? Use --dry-run:
ssg init --dry-run --client=claude-code
ssg init prints a per-file preview showing exactly which files and lines it will create or modify, with revert commands next to each one. Confirm with Y to proceed.
After install, .sigmashake/install-manifest.json records every file SSG touched. Re-read it any time with ssg whatchanged. For the full breakdown of every config file SSG knows how to write, see What SSG Changes on Your Machine.
This writes the ssg-governance MCP server into your AI client's config file — no manual JSON editing required. Restart your AI client once.
Step 3 — Just ask
Open your AI agent and say:
"Set up SigmaShake governance for my project"
The agent calls ssg_onboard, which automatically:
- Detects your tech stack (TypeScript, React, Go, Python, Docker, etc.)
- Creates
.sigmashake/config.tomland starter security rules - Pulls matching rulesets from hub.sigmashake.com
- Runs
ssg lintto verify everything - Returns a health summary
That's it. You're protected.
What else can you ask?
| Say this to your agent | What happens |
|---|---|
| "Set up SigmaShake for my project" | Full one-shot setup (ssg_onboard) |
| "Check if governance is working" | Health diagnostics (ssg_doctor) |
| "Find Docker security rules on the Hub" | Ruleset search (ssg_hub_search) |
| "Install rules-typescript" | Ruleset installation (ssg_hub_pull) |
| "Block any npm publish without my approval" | Custom rule creation (ssg_write_rule) |
| "Show my current plan and active rules" | Status report (ssg_status) |
Option B — Manual CLI Setup
Prefer to do it yourself? Subscribe at sigmashake.com/pricing, then run these commands:
# 1. Authenticate (browser opens accounts.sigmashake.com)
ssg auth login
# 2. Initialize your project
cd your-project
ssg init
# First run: you will be prompted to accept the SigmaShake Terms of Service
# (https://sigmashake.com/terms) before the dashboard is available.
# In CI or non-TTY environments, pass --accept-terms to proceed non-interactively.
# The data directory defaults to ~/.sigmashake; use --home <dir> or SIGMASHAKE_HOME
# to relocate it.
# 3. Pull rulesets for your stack
ssg hub pull rules-typescript # adjust to your stack
ssg hub pull rules-react
ssg hub pull rules-docker
Verify everything works:
ssg lint # check rule syntax
ssg status # system health + active rule count
ssg doctor # full diagnostic report
Start the governance dashboard (optional — shows real-time audit log + approval queue):
ssg serve # opens http://localhost:5599
Authentication
Authentication links your machine to your SigmaShake plan tier (Pro / Enterprise).
Browser login (recommended)
ssg auth login
Select Browser (recommended). Your browser opens to accounts.sigmashake.com — if you're already signed in via GitHub, Google, or your company's SSO, just click Authorize ssg CLI. Token flows back to the CLI automatically.
Enterprise SSO
ssg auth login --sso
See Fleet SSO Setup for Okta, Azure AD, and SAML 2.0 guides.
API Key
ssg auth login --token=YOUR_API_KEY
Write your first rule
Rules live in .sigmashake/rules/. Create a file and write:
rule block-force-push {
enabled true
priority 100
severity error
DENY execution
IF command CONTAINS "git push --force"
OR command CONTAINS "git push -f"
MESSAGE "Force push is blocked. Use --force-with-lease instead."
}
Then verify:
ssg lint # ✓ 1 rule OK
ssg list # shows all active rules
See Rule Syntax for the full language reference.
Test a rule manually
echo '{"tool":"Bash","input":{"command":"git push --force origin main"}}' | ssg eval
Output:
{
"decision": "block",
"rule_id": "block-force-push",
"reason": "Force push is blocked. Use --force-with-lease instead.",
"duration_ns": 1840000
}
Integrate with Claude Code
ssg init --client claude-code
This patches .claude/settings.json with:
- Hook —
ssg hook evalruns before every tool call - Permissions — 14 standard tools pre-approved, 18 destructive patterns pre-denied
See Claude Code Integration for the full setup guide.
Next steps
- MCP Server — Complete tool reference for AI agents
- Rule Syntax — Full rule language reference
- Writing Rules — Best practices and patterns
- Hub — Browse and install community rulesets
- CLI Reference — All 30+ commands
- Enterprise Fleet — Fleet admin onboarding via AI agent
Join the Community
Get help, share your rulesets, and follow what's shipping: