SuperClawd brings your team's coding standards, skills, and guardrails into every Claude Code session — delivered on demand, with no context bloat. Define your rules once and Claude Code follows them.

How does SuperClawd work?

Three steps to consistent AI output. First, define your rules by creating instructions for your workspace's standards, guidelines, and best practices, organized by custom skills you define. Second, run superclawd to launch Claude Code with your workspace's skills, agents, commands, and rules already loaded — no manual setup. Third, every session applies your workspace guidelines automatically, with the right instructions delivered on demand exactly when they're needed.

What features does SuperClawd offer?

SuperClawd provides a suite of tools: Skills group your instructions into logical sets such as coding standards, writing styles, or domain expertise; Workflows let you design multi-step processes with a visual editor; Commands create custom slash commands for quick actions; and Resources store reusable templates, code snippets, and reference materials that AI pulls on demand. Versioning, experimental mode, and share links work the same way across all four — skills, agents, commands, and workflows — on a shared internal core: ship each as named, immutable releases, promote a stable version for the whole team, roll back instantly, restore any release to keep iterating, and test drafts behind a two-key experimental flag. Unreleased entities are never served. Unlisted, revocable share links let you hand any of them to another workspace, where recipients preview and import their own copy. Workspace memories inject shared team facts into every member's Claude Code session at startup. Workspace MCP management lets you curate the MCP servers your whole team can use — stdio or HTTP — and the CLI injects the enabled ones into every member's session automatically, with secrets encrypted at rest and one-click presets like Linear and Context7.

How much does SuperClawd cost?

SuperClawd is pay as you go with no subscriptions. The Free plan is $0 with limited usage. The Starter plan is $10 for 100 credits. The Growth plan is $50 for 550 credits, including a 10% bonus of 50 extra credits. More packages are available in the app, and auto top-up is available.

Why use SuperClawd instead of plain Claude Code?

Without SuperClawd, AI guesses your preferences, produces inconsistent output, and requires manual corrections every time while wasting tokens on repeated context. With SuperClawd, the AI knows your rules and produces consistent output every time, gets it right the first time, uses fewer tokens through on-demand instructions, and notifies you when it needs your input.

Does SuperClawd offer support?

Yes. SuperClawd provides in-app support from real humans on our team — no chatbots. You can also join the community discussion on GitHub.

SuperClawd | Claude Code, supercharged

Everything you need

Powerful tools to control how AI works for you.

Skills

Group your instructions into logical skills — structured categories or free-form markdown. Capture coding standards, writing styles, or domain expertise, applied contextually.

Agents

Author subagent or primary personas with their own system prompt, per-tool permissions, and an optional step cap — then version and share them like everything else.

Teams

Compose a lead agent with member agents and linked workflows. Point a command at the team and one prompt spins up a coordinated crew — right inside Claude Code.

Workflows

Design multi-step processes with a visual editor. Connect nodes to create decision trees, checklists, and automated sequences.

Commands

Create custom slash commands for quick actions. Type /review or /deploy and let AI execute predefined instructions.

Resources

Store reusable templates, code snippets, and reference materials. AI pulls them on-demand to provide consistent, accurate context.

Versioning & Sharing

One release model across skills, agents, commands, workflows, and guardrails — immutable releases, a promotable stable version, instant rollback, an experimental flag, and revocable share links.

Team Memories

Inject shared team facts into every member's Claude Code session at startup. Set standards once — they're there from the first prompt, while a member's explicit request always wins.

Learned Memory

Your AI captures durable facts as you work — repo, agent, and private tiers — and recalls them next session. Captured locally with zero server-side LLM; private stays yours.

MCP Management

Curate the MCP servers your team can use — stdio or HTTP — and the CLI injects them into every session automatically. Secrets stay encrypted and no one hand-edits config.

Guardrails

Admin-defined rules that intercept tool calls via PreToolUse — Deny blocks, Ask prompts for approval, Warn notes and proceeds. Enforced team-wide, no per-developer setup.

Pipelines

Pipelines is a local web app that runs your own Claude on your tickets. Assign an issue from Linear, Jira, Asana, ClickUp, or Monday and Claude plans it, works in an isolated git worktree, opens a PR, and hands it back for your review — everything on your machine, launched from the CLI.

How it Works

Define your standards once — Claude follows them every session.

Define

Create instructions for your workspace's standards, guidelines, and best practices. Organize by custom skills you define — no restrictions.

Launch

Run superclawd to start Claude Code with your workspace's skills, agents, commands, and rules already loaded — no manual setup.

Apply

Every session applies your workspace guidelines automatically — the right instructions delivered on demand, exactly when they're needed.

Turn your board into an autonomous pipeline

Assign the work and it gets done — scaling from a single project to your entire roadmap.

Assign

Pick up an issue from Linear, Jira, Asana, ClickUp, Monday, or a built-in local board.

Plan

Your own Claude reads the ticket, plans the work, and gets going — supervised or autonomous, per run.

Work

Each ticket works in its own git worktree and opens its own PR, so concurrent runs never collide.

Review

Claude hands the PR back for your review, with a human sign-off before anything is marked Done.

Works your real board — Linear, Jira, Asana, ClickUp, Monday, or a built-in local board.
Autonomous runs, isolated by design — each ticket works in its own git worktree and opens its own PR; concurrent runs never collide.
You stay in control — supervised or autonomous per run, with a human sign-off before anything is marked Done.
Recurring schedules + a PR monitor — run work on a cadence, and auto-fix failing CI on open PRs.
Local-first — your data lives on your machine; nothing runs in our cloud.

superclawd

$ superclawd pipelines start |

✓ Pipelines running at http://localhost:63210
Assigned ENG-482 → planning…
Working in isolated worktree .worktrees/eng-482
✓ Opened PR #128 — ready for your review

Requires the SuperClawd CLI and SuperClawd Pipelines.

Docs Read the announcement

Claude Code, supercharged.

The difference is clear

Without SuperClawd

With SuperClawd

Everything you need

Skills

Agents

Teams

Workflows

Commands

Resources

Versioning & Sharing

Team Memories

Learned Memory

MCP Management

Guardrails

Pipelines

How it Works

Define

Launch

Apply

Turn your board into an autonomous pipeline

Assign

Plan

Work

Review

Pricing

Free

Growth

Starter

Real humans. No chatbots.

Claude Code, supercharged.

The difference is clear

Without SuperClawd

With SuperClawd

Everything you need

Skills

Agents

Teams

Workflows

Commands

Resources

Versioning & Sharing

Team Memories

Learned Memory

MCP Management

Guardrails

Pipelines

How it Works

Define

Launch

Apply

Turn your board into an autonomous pipeline

Assign

Plan

Work

Review

Pricing

Free

Growth

Starter

Real humans. No chatbots.

Stay Updated

Stay Updated