SG Agent Harness V2 • Deterministic Multi-Agent Verification

The Deterministic Verification & Context Harness for AI Coding Agents

AI agents recommend. Deterministic gates decide. Scaffold, govern, and verify multi-agent development for Claude Code, Cursor, Codex, Antigravity, and OpenCode with zero hallucinated completion.

16
Verification Gates
5
Multi-Agent Adapters
29+
Progressive Skills
100%
Deterministic Evidence
Published successfullysg-harness@2.0.0 is live 🎉

One-command setup for Claude Code, Cursor, Codex, Antigravity & OpenCode

View on npm
Terminal — Initialize Project
# Navigate to your project directory and initialize
cd /path/to/your/project
npx sg-harness init --yes
# Creates AGENTS.md, thin adapters, MCP configs, and local ./sg CLI
Interactive Control Center

Agent Harness Generator

Curated Harness Presets

One-Click Production Stacks

Select a preset or customize below
npx @sg/harness init my-saas-app \
  --framework=nextjs \
  --mode=PRODUCTION \
  --agents=claude-code,cursor,antigravity \
  --domains=frontend,backend,database,api-contracts,auth \
  --context=graphify,hydradb,context7,framework-docs
Configuration: NEXTJS • PRODUCTION100% Deterministic
Architecture & Workflow

The 16-Step GAUNTLET Verification Loop

GAUNTLET mode runs an adversarial zero-trust review loop across 8 critic roles, ensuring that no feature can be certified until all measurable acceptance criteria and fault injections are defended.

01 Understand
02 Research
03 Graphify
04 Memory
05 Plan
06 Implement
07 Unit Tests
08 Browser E2E
09 Visual QA
10 A11y AA
11 Security
12 Performance
13 Adversarial
14 Fix Loop
15 Re-Test
16 Certified
Progressive Skill Registry

Active Agent Skills Catalog

29 specialized skills routed via .sg/routing.json for maximum relevant context.

Last 30 Days Researchresearch

User requests real-world consensus, recent 2026 opinions, or discussions from Reddit/HN/X/GitHub.

last30days/SKILL.mdKEEP
Context7 Documentation Searchresearch

Working with an unfamiliar API, third-party SDK, or recently updated library syntax.

find-docs/SKILL.mdKEEP
Socratic Requirement Interviewplanning

User asks for interactive requirement discovery, product scoping, or trade-off evaluation before execution.

grill-me/SKILL.mdKEEP
Architectural Ideationplanning

Initial phase of open-ended features requiring multiple structural variations or trade-offs.

brainstorming/SKILL.mdKEEP
Graphify Code Intelligencearchitecture

Understanding codebase structure, dependency paths, AST nodes, or blast radius before refactoring.

graphify/SKILL.mdKEEP
System Architecture Refactoringarchitecture

Major refactoring, structural decoupling, code bloat elimination, or harness evolution.

improve-codebase-architecture/SKILL.mdKEEP
Modern Frontend Engineeringfrontend

Building or refactoring React, Astro, or TypeScript UI components, islands, and layouts.

frontend-dev/SKILL.mdKEEP
Tailwind CSS v4 Enginefrontend

Writing CSS tokens, @theme definitions, dynamic utilities, or responsive layout styles.

tailwind-4-docs/SKILL.mdKEEP
Emil Kowalski Design Craftdesign

Polishing UI typography, micro-details, optical alignment, layout rhythm, and interactive states.

emil-design-eng/SKILL.mdKEEP
Anti-Generic Visual Designdesign

Creating visually stunning pages, landing screens, color schemes, and avoiding generic AI templates.

taste-skill/SKILL.mdKEEP
Motion Physics & Tuningmotion

Tuning spring curves, damping, layout transitions, exit/enter animations, or hover micro-interactions.

review-animations/SKILL.mdKEEP
GSAP ScrollTrigger & Pinningmotion

Complex scroll-driven parallax, scrubbing, timeline sequencing, or canvas animations.

gsap-scrolltrigger/SKILL.mdKEEP
Go Backend Context & Concurrencybackend

Writing Go services, goroutines, cancellation contexts, or backend APIs.

golang-context/SKILL.mdKEEP
Rust Systems & Memory Safetybackend

Writing Rust services, CLI tools, async tokio runtimes, or WASM modules.

rust-best-practices/SKILL.mdKEEP
HydraDB Durable Project Memorydatabase

Querying or recording architectural decisions, ADR conclusions, rejected options, or debugging discoveries.

.sg/scripts/memory.tsKEEP
Zero-Fabrication ValidationAI

Working with LLM prompt engineering, structured extraction, JSON rescue, or resume parsing.

no-mistakes/SKILL.mdKEEP
Security & PR Verification Gatesecurity

Verifying code against security rules, OWASP Top 10, secrets leaks, and input sanitization.

pr-feedback-quality-gate/SKILL.mdKEEP
Browser Subagent Inspectiontesting

Visual UI rendering verification, end-to-end user flows, layout checks, or DOM inspection.

agent-browser/SKILL.mdKEEP
Client Download & Export Debuggingdebugging

Debugging blob creation, file downloads, PDF generation, or clipboard copy failures.

export-download-debugging/SKILL.mdKEEP
Release Notes & Changelog Generatorrelease

Preparing production release summaries, git tag changelogs, or customer-facing updates.

release-notes-one-pager/SKILL.mdKEEP
Visual QA & Design Criticreview

Substantial UI changes, responsive layout refactors, design-system tokens, theme parity, or GAUNTLET mode visual critiques.

sg-visual-qa/SKILL.mdKEEP
Accessibility & WCAG AA Enforcementaccessibility

Creating or modifying UI components, forms, navigation dialogs, color themes, or running PRODUCTION/GAUNTLET gates.

sg-accessibility/SKILL.mdKEEP
Multi-Tenant Authorization & IDOR Reviewsecurity

Implementing auth flows, database models, RLS policies, team memberships, or API endpoints handling tenant resources.

sg-auth-review/SKILL.mdKEEP
Production Security & SAST Scannersecurity

Performing security audits, dependency vulnerability checks, injection defense, or secret leak protection.

sg-security-review/SKILL.mdKEEP
Supabase & Postgres Best Practicesdatabase

Creating/altering tables, writing SQL migrations, configuring RLS, adding indexes, query tuning with EXPLAIN ANALYZE, or pgvector similarity search.

supabase-postgres-best-practices/SKILL.mdKEEP
Zero-Downtime Migration Reviewerdatabase

Writing or reviewing schema migrations, table locks, column additions, batched backfills, or rollback plans.

sg-db-migration-review/SKILL.mdKEEP
Web Performance & Dependency Gatekeeperperformance

Optimizing Core Web Vitals (LCP, CLS, INP), evaluating asset bundles, auditing dependencies, or tuning hydration.

sg-perf-review/SKILL.mdKEEP
Production Observability & Sentry Telemetrydebugging

Configuring Sentry, structured logging, request correlation IDs, error boundaries, or health endpoints.

sg-observability/SKILL.mdKEEP
Zero-Trust Adversarial GAUNTLET Modereview

Major features, architecture shifts, security changes, pre-release certification, or explicit gauntlet invocation.

sg-gauntlet/SKILL.mdKEEP
Universal CLI Dispatcher

SG Command Cheatsheet

Universal multi-agent command interface for humans, Claude Code, Cursor, Codex, Antigravity, OpenCode, and CI.

sg-terminal
$./sg verify

Dynamically resolves applicable gates and executes deterministic checks.

Simulated Output
✔ VERIFIED (PASSED) - 8/8 gates passed (.sg/evidence/latest.json)
Deterministic Zero-Trust ProtocolSG Agent Harness V2
Thin Adapter Architecture

Multi-Agent Compatibility Matrix

Single canonical source of truth in AGENTS.md. Thin adapters translate into each harness natively without rule drift.

Coding AgentNative ConfigAGENTS.md SupportSkill RoutingMCP IntegrationCompletion Gate
Claude CodeAnthropicCLAUDE.mdFull Root PointerVia .sg/routing.json.claude/mcp.json./sg verify
OpenAI CodexOpenAICODEX.mdFull Root PointerProgressive disclosureTool-calling protocol./sg verify
Cursor IDECursor / Anysphere.cursor/rules/*.mdcFull MDC PointerRule Triggers.cursor/mcp.json./sg verify
Google AntigravityGoogle DeepMind.agents/AGENTS.mdNative KernelNative .agents/skills/Native eager/lazy./sg verify
OpenCode / OpenHandsOpen Sourceopencode.jsoninstructions fieldVia schemaStandard MCP./sg verify