.claude

Creates a new specialist agent as two stitched source files, base.md plus frontmatter/claude.json, proposing model, effort, and permissions by role archetype and confirming them with the user before writing. Use when adding a new subagent, defining a new specialist role, scaffolding an agent definition, or when update-agent hands off new-agent creation.

Discover, validate, stitch, and install specialist agent templates contributed by Essential and other enabled plugins in the same marketplace. Use when asked to install agents, set up subagents, refresh the agent team, or configure Claude Code on a new machine.

[Description] — third person, what the skill does plus a 'Use when...' clause with concrete trigger phrases a user would actually type; 25-60 words; lead with the key use case.

Update explicitly selected agent definitions to the current two-file template or a stated behavior change while preserving useful role expertise, trigger boundaries, context, collaboration links, and working voice. Use when migrating agents to a template revision, correcting agent configuration, or batch-updating selected agents; require an exact selector and route genuinely new roles to create-agent.

Enforce coding standards mechanically across a selected scope with batched linters and independent reviewers. Use when source files need lint-error correction, standards enforcement, or consistent formatting, including calls extended by another plugin's portable lint profile; behavior-changing repairs belong to fix.

Review semantic correctness, security, test intent, documentation, sibling consistency, and alignment with the implementation plan. Use after code changes or for explicit review requests; report findings without editing code and leave mechanical standards enforcement to lint.

Materialize a guaranteed-on-disk Notion specification tree as a flat `.code-spec/` bundle of `{kebab-title}-{32hex-id}.md` files plus `.gitignore`. Use before downstream analysis or code generation, when refreshing a stale bundle, or when a ticket requires local spec evidence; fail unless the root id-suffix file exists and is non-empty.

Design or redesign a web interface — and implement it when authorized — with coherent visual direction, responsive layout, typography, color, motion, and accessible states. Maintains a .design task workspace and ranked variant boards, then drives an independent implement-evaluate loop with visual-diff confirmation. Use for new pages, component polish, mockups, or facelifts.

Diagnose Next.js runtime behavior with next-browser and Chrome DevTools MCP: React components, routes, SSR errors, DOM/styles, performance, Lighthouse, network, device emulation, JavaScript debugging, storage, screenshots, and interactions. Use for evidence-backed browser diagnosis; route visual creation to design and story-state assessment to storybook.

Persist CONTEXT.md, NOTES.md, and PLAN.md for later continuation of coding work. Use when pausing implementation or transferring repository state; this skill records the current session and does not create or execute a cross-domain plan.

Resume paused coding work from CONTEXT.md, NOTES.md, and PLAN.md. Use for takeover, continuing yesterday's coding work, resuming a continuation bundle, or --revalidate; trust recorded assumptions for 24 hours, revalidate older state, and invoke coding:write-code --resume. For saving current state, use the session-persistence workflow.

Write production-ready code end to end through a TDD lifecycle of design, skeleton, implementation, tests, and refactoring. Use for new functions, features, modules, components, CLI or API endpoints, or approved tickets; route diagnosed failures to fix and explicit production stubs to complete-code.

Decides between researched approaches before implementation. Use when asked to choose an approach, challenge a recommendation, make an architecture decision, compare options, define rollback and falsification signals, or obtain approval; routes blindspot passes, brainstorms, interviews, references, and prototypes to essential:discover.

Discovers material unknowns before planning. Use for a blindspot pass or unknown unknowns, to brainstorm approaches from cheapest to ambitious, interview about architecture, extract reference implementation semantics, make a disposable prototype before touching the real app, or check whether discovery is ready for a decision; researched option selection belongs to essential:decide.

Create or execute a context-complete cross-domain plan as an orchestrator. Use when another agent must continue without prior context, or when a multi-domain plan needs coordinated execution while this skill retains decision ownership. For coding-session persistence, use coding:handover.

Execute an approved specification ticket from authoritative contract through implementation, review, and commit planning. Use after plan-code approval, when resuming partial ticket work, or when auditing a delivered ticket. Keep contract authoring in spec-code and generic feature work in coding:write-code.

Generate DRAFT.md as a commit blueprint and PLAN.md as an execution roadmap from an approved proposal or specification. Use when planning implementations, defining atomic commits, documenting change proposals, or preparing a coding workflow with explicit verification and ownership boundaries.

Design or document technical specifications in the canonical template, then delegate Notion synchronization to sync-notion. Use for greenfield specs, updates to an existing DESIGN.md, or documenting an implementation without inventing requirements.

Use when validating a new or changed Claude Code skill, checking structural and repository policy compliance, testing whether descriptions trigger accurately, or grading representative skill outputs before deployment.

Audit a rendered web interface against the design standard with the bundled deterministic CLI, shared-browser evidence, responsive viewports, accessibility checks, and focused visual adjudication. Use for design QA, WCAG checks, visual review, or launch assessment. Produce reports and evidence only; route fixes to the owning implementation skill.

Author and improve tests for pending test cases, coverage gaps, fixtures, and redundancy cleanup. Use for test TODOs, it.todo or describe.todo entries, explicit test-writing requests, or coverage work. Production implementation stubs belong to complete-code; diagnosed failures belong to fix.

Use when creating a reusable Claude Code skill, defining a new repeatable agent capability, or replacing a one-off workflow with discoverable instructions that need clear ownership, validation, and trigger behavior.

Use when revising one or more existing Claude Code skills, aligning skill instructions with current repository policy, narrowing overlapping ownership, or applying a deliberate behavior change without creating a competing skill.

Merge a linear stack of GitHub pull requests while restacking descendants between merges. Use when a user invokes /merge-pr with PR numbers, asks to merge stacked PRs, or needs gh-driven bottom-up PR merging with automatic downstream rebase.

Save code changes cleanly with jj-first, git-compatible routing. Use for commits, split/absorb/edit operations, stacked changes, history reordering, retrospective blame fixes, or the --create-pr compatibility handoff; preserve the repository history policy and keep coding:commit as the sole history-mutation owner.

Run isolated per-commit QA across every unpushed commit, report ordering or message issues, and coordinate approved corrections. Use before publishing a stack; coding:commit owns history mutations and coding:push-pr owns publication.

Publish saved changes as draft pull requests and drive GitHub CI to green. Use when asked to push the latest commit, create or update a PR, repush after a fix, babysit pending checks, repair red CI, monitor every check, or converge a PR stack.

Install or update registered coding CLI tools (brew, jj, gh, fallow, python) across macOS, Linux, and Windows. Use when tools are missing, stale, or needed on PATH for a sibling skill, including requests to install jj/gh/brew, update coding tools, or verify CLI dependencies before work.

Author a conventional-commit PR title and unified body from a jj or git change ref, emitting output for gh pr create. Use for PR descriptions, draft pull requests, stacked coding:push-pr PR bodies, and callers that need a unified title/body template from a commit.

Audit stale development state across git branches, remote branches, git worktrees, and jj workspaces. Use when a user asks to run /cleanup, find abandoned branches/workspaces, or safely remove already-merged or duplicate local work with backups and confirmation.

Scaffold or maintain a project's root stylesheet using the CSS-only light, dark, and system color-mode contract. Use for theme.css, globals.css, or app.css setup, migration from class-driven dark mode, semantic token wiring, or color-mode corrections. Detect conflicts, obtain migration approval, preserve existing tokens, and edit CSS only.

Generate or edit images through the bundled multi-provider CLI, or write structured prompts and analyze visual styles from references. Use for concept art, product shots, covers, UI assets, transparent or vector output, inpainting, background changes, batch variants, and prompt-only work. Keep image generation separate from web design decisions and visual audits.

Audit a Storybook instance for setup failures, accessibility violations, interaction errors, and visual regressions across meaningful story states. Use before release or when validating addons and focus behavior. Run the bundled lifecycle in order, preserve evidence, and report findings; do not edit components, stories, or configuration.

Create a new technical standard at a plugin's canonical constitution/standards root using meta.md, scan.md, write.md, and per-rule guides. Use when establishing new coding standards, documenting technical requirements, or creating compliance guidelines for reusable policy with explicit dependencies, detection, compliant patterns, and stable rule IDs. Route existing-standard revisions to update-standard.

Update explicitly selected plugin standards to the current meta.md, scan.md, write.md, and rules contract while preserving valid policy and stable rule IDs. Use when applying scoped rule changes, migrating standards to a template revision, or batch-updating the standards library. Require a path, glob, or --all; route missing targets to create-standard.

Use when React JSX, components, hooks, accessibility, project structure, tests, or Storybook files need mechanical standards enforcement through the shared Coding lint workflow; React owns framework rules while Coding owns generic execution and reporting.

Use when creating, editing, reviewing, or routing work involving React, JSX, hooks, components, accessibility behavior, project structure, tests, or Storybook stories; this router selects React standards while Coding owns generic execution.

Read, edit, and author MDC (Contextual Markdown, @theriety/mdc) files safely with native text tools. Use when asked to "edit this .mdc file", "add a block to <doc>.mdc", "update the annotation for ref <x>", "convert this to MDC", whenever a .mdc file must be read or written, or when mutating any file under .code-spec/.

Review an implementation against an authoritative local or Notion specification, then run the general coding and security review. Use for specification alignment, delivered-ticket validation, or detecting omissions, drift, and unsanctioned behavior before handoff.

Synchronize one or more paired Markdown files and Notion pages in a declared direction. Use when local documentation must be published, remote pages must be materialized locally, or both sides require an explicit conflict-resolved merge. Keep specification authoring in spec-code.