carbon

Agentically test a specific feature by analyzing the branch diff, building a test plan, and driving the app through agent-browser. Builds on /login and /error skills. Caches successful playbooks to .ai/playbooks/ for reuse.

Research-driven brainstorming and design for a feature. Use when designing a new feature, exploring approaches, or making design decisions. Automatically researches competitors before asking design questions, then hands off to /plan for implementation. Triggers on "brainstorm", "design", "design a feature", "plan a feature", "how should we build X", or any request to explore approaches for a feature.

Author, edit, or extend the Carbon documentation site at `docs` (a built Fumadocs + Next.js app). Use whenever creating or changing reader-facing docs for Carbon: editorial Guide chapters, Reference/entity pages, the docs IA, or "document this feature" requests — anything destined for the docs site. Covers the grounded-in-source authoring workflow, the flow-based Guide architecture, the real MDX components for each surface, the warm-paper house style, and the build verification loop. Trigger even if the user doesn't say "docs" but clearly wants explanatory, reader-facing written material about how Carbon works.

Supervised loop conductor — drive a single work item (a bug, a usability tweak, a small feature) through a doer→gate→keep-or-revert→ledger cycle to a gated PR, while the human watches. Use when asked to "conduct", "build", "fix", or "loop on" a specific item with the conductor. Supervised only — not autonomous/overnight (not yet supported). Backed by @carbon/harness + the @carbon/checks checks.

Execute implementation plans task by task with verification. Use after /plan to implement the feature. Follows each task exactly, runs verifications, commits frequently. Triggers on "execute the plan", "implement the plan", "run the plan", or after /plan approval.

End-to-end feature development orchestrator. Composes /research, /brainstorm, /plan, and /execute into a single workflow. Use when building a new feature from scratch. Triggers on "build a feature", "implement X from scratch", "full feature development for Y".

Survey the Carbon codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on source code — never implements, fixes, or refactors anything itself. Use when asked to audit Carbon, find improvement opportunities (bugs, security, performance, test coverage, tech debt, migrations, DX), suggest features or where to take the project next (roadmap, product direction), or generate handoff plans for another agent to implement.

Create detailed implementation plans from design specs. Use after brainstorming to create step-by-step implementation plans. Each task is 2-5 minutes of work with exact code and commands. Triggers on "plan the implementation", "create a plan for", "write the implementation plan", or after /brainstorm approval.

Survey best practices from competitor ERPs and point solutions for a given feature. Use when designing new features, evaluating approaches, or understanding industry patterns. Triggers on "research best practices", "how do competitors do X", "survey the market for Y", "what's the industry standard for", or any request to understand how other systems handle a feature.

Critically review your own PR work before or just after opening the PR. Use when finishing a branch, before opening or merging a PR, or to sanity-check a diff against main. Produces Must fix / Risks / Suggested improvements. Supports an opt-in strict "thermo-nuclear" mode for a deep maintainability and abstraction audit when explicitly requested.

Use when creating new skills, editing existing skills, or verifying skills work before deployment