Turns Claude Code into a full product-development toolkit: an opinionated plan → verify → commit engineering workflow, specialized subagents for legal compliance, UX design, backend/frontend/cloud architecture, Go/Java/TypeScript implementation, and DevOps/infrastructure (Terraform, Kubernetes, Docker, AWS/Azure/GCP), plus a code reviewer that enforces your language conventions automatically. Subagents are namespaced — invoke as @addit-harness:code-reviewer, not @code-reviewer. Curated and adapted from established open-source sources, not hand-rolled from scratch.
Use when the user makes or records a significant architectural or technical decision — choosing a framework, database, protocol, pattern, or trade-off ("we decided to use X instead of Y because..."), or explicitly asks to write an ADR. Captures the decision as a structured MADR-format Architecture Decision Record in docs/adr/.
Scan the current TS/React project's existing UI layer and write .claude/design-conventions.md — a project-specific visual design convention file that rules/typescript.md loads before UI work. Use when starting UI work in an existing project, when design conventions have drifted, or when --refresh is passed to merge in new patterns.
Scan the current Go repo and write .claude/go-conventions.md — a project-specific convention file that rules/go.md loads on every session instead of re-scanning the codebase from scratch. Use when starting work in a new Go repo, when conventions have drifted, or when --refresh is passed to merge in new patterns.
Use when the user wants to save, export, or persist a plan (or any diagram-rich design doc shown in the CLI) to a markdown file so it can be opened and viewed with rendered mermaid diagrams. The terminal can't render mermaid; this writes the plan to docs/plans/ (or a temp scratch file) for viewing in an IDE preview or on GitHub.
Use once, right after installing the addit-harness Claude Code plugin, to place the parts a plugin can't carry natively (CLAUDE.md, AGENTS.md, path-scoped rules/, references/, settings.json) into ~/.claude or the current project. Also use to re-sync after a plugin update, or to switch between global and project scope.
Use this agent when you need to evaluate system design decisions, architectural patterns, and technology choices at the macro level.
Expert backend architect specializing in scalable API design, microservices architecture, and distributed systems. Masters REST/GraphQL/gRPC APIs, event-driven architectures, service mesh patterns, and modern backend frameworks. Handles service boundary definition, inter-service communication, resilience patterns, and observability. Use for up-front DESIGN of new backend services or APIs (contracts, boundaries, trade-offs) — produces design docs and guidance, not the implementation. To WRITE the code, use backend-developer (or frontend-developer for UI).
Use this agent to IMPLEMENT backend features and fixes in code once the design/approach is clear. A senior backend engineer and software craftsman that writes, structures, tests, and verifies code in Go, Java/Spring, and TypeScript/Bun, following this setup's vendored conventions. Use PROACTIVELY for backend implementation/coding tasks. For up-front architecture/API design use backend-architect; for UI work use frontend-developer; for review use code-reviewer.
Expert cloud architect specializing in AWS/Azure/GCP/OCI/DigitalOcean multi-cloud infrastructure design, advanced IaC (Terraform/OpenTofu/CDK), FinOps cost optimization, and modern architectural patterns. Masters serverless, microservices, Kubernetes, edge/network security (Cloudflare WAF/DDoS/Zero Trust and hyperscaler-native equivalents), cloud security posture (GuardDuty/Security Hub, Defender for Cloud, Security Command Center), compliance, and disaster recovery. Handles both up-front DESIGN of new infrastructure and REVIEW of existing infrastructure (cost, security, reliability audits). Use PROACTIVELY for cloud architecture, cost optimization, migration planning, multi-cloud strategies, or auditing existing infra. To WRITE and VERIFY the actual Terraform/Kubernetes/Docker code, use devops-engineer.
Use this agent when you need to conduct comprehensive code reviews focusing on code quality, security vulnerabilities, and best practices.
Use when the user makes or records a significant architectural or technical decision — choosing a framework, database, protocol, pattern, or trade-off ("we decided to use X instead of Y because..."), or explicitly asks to write an ADR. Captures the decision as a structured MADR-format Architecture Decision Record in docs/adr/.
Scan the current TS/React project's existing UI layer and write .claude/design-conventions.md — a project-specific visual design convention file that rules/typescript.md loads before UI work. Use when starting UI work in an existing project, when design conventions have drifted, or when --refresh is passed to merge in new patterns.
Scan the current Go repo and write .claude/go-conventions.md — a project-specific convention file that rules/go.md loads on every session instead of re-scanning the codebase from scratch. Use when starting work in a new Go repo, when conventions have drifted, or when --refresh is passed to merge in new patterns.
Use when the user wants to save, export, or persist a plan (or any diagram-rich design doc shown in the CLI) to a markdown file so it can be opened and viewed with rendered mermaid diagrams. The terminal can't render mermaid; this writes the plan to docs/plans/ (or a temp scratch file) for viewing in an IDE preview or on GitHub.
Use once, right after installing the addit-harness Claude Code plugin, to place the parts a plugin can't carry natively (CLAUDE.md, AGENTS.md, path-scoped rules/, references/, settings.json) into ~/.claude or the current project. Also use to re-sync after a plugin update, or to switch between global and project scope.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A coding agent's out-of-the-box config is a blank slate. addit-harness is the config harness that turns it into a full digital product development kit — covering every layer from idea to ship: legal compliance, UX design, architecture, implementation in Go · Java · TypeScript, and code review, all wired into a plan→verify→commit engineering loop. One-line install, auto-detecting whichever agent(s) you have. Philosophy: curate established assets and adapt them — don't hand-roll what already exists.
A shared configuration that covers the full digital product development loop — from idea investigation and legal compliance through UX design, backend/frontend architecture, and implementation in Go · Java/Spring · TypeScript (Bun backend · React/Next.js frontend), to code review and debugging — synced into Claude Code, Cursor, Codex CLI, and Kiro. See Other coding agents for what's covered per tool (GitHub Copilot support is planned, not yet implemented — see Roadmap). Built for SaaS founders and product engineers who own the full feature lifecycle and want every layer of that loop to have an opinionated, specialized subagent behind it.
Philosophy: adopt established, well-known assets (the official plugin
marketplace + reputable community collections) and adapt/pin them — don't
hand-roll what already exists. The repo is a thin curation + config layer, not
a pile of bespoke skills. It deliberately reuses Claude Code's built-ins
(/code-review, /simplify, /verify, /run, /init, deep-research).
The condensed, browsable version of this README lives at
tools.addit.digital/harness/docs/
— grouped sidebar, on-page tables of contents, and per-tool install pages.
Prefer browsing in-repo instead? Start at
site/_docs/README.md.
Without it: Claude Code starts as a blank slate — no engineering loop, no conventions, no subagents. You either wing it session to session or spend hours wiring up memory, rules, and delegation yourself, then rebuild it on every machine.
With it: one command gives you a reproducible, opinionated starting point:
| Blank slate | This config | |
|---|---|---|
| Engineering process | Ad-hoc | plan→verify→commit loop baked in |
| Legal compliance | Manual check (or skipped) before shipping | @saas-legal-advisor assesses impact of every feature change; drafts/reviews T&C, Privacy Policy, cookie policies, DPAs |
| Language conventions | Manual context injection every session | Auto-load per file type (Go · Java · TS) |
| Code review | Ask Claude to review | @code-reviewer enforces the conventions with file:line citations |
| Complex work | Monolithic context | Delegate to specialized subagents; main context stays clean |
| UX & design | Describe and hope | @ux-designer → @figma-designer → @frontend-architect pipeline |
| New machine | Redo everything | ./install.sh |
No clone, no shell script — install the plugin from inside Claude Code:
/plugin marketplace add addit-digital/addit-harness
/plugin install addit-harness@addit
/addit-harness:setup
npx claudepluginhub addit-digital/addit-harness --plugin addit-harnessComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex, antigravity, and grok CLIs when installed) to get diverse perspectives on coding problems
Harness-native ECC plugin for engineering teams - 67 agents, 279 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Lazy senior dev mode. Forces the simplest, shortest solution that actually works: YAGNI, stdlib first, no unrequested abstractions.