nilsclabb/yellowpages

pr-code-review

Specialized review leaf for PR, branch, and diff review. Use from `yp-workflow-review` when the user explicitly wants PR-oriented review or merge-readiness analysis.

yellowpages

The yellowpages skill defines how agents should create, author, and package skills, structure skills for navigation, navigate between skills and documentation, document planning work, define agent personas, run and track multi-step workflows, track state…

using-yellowpages

Use when starting a session with yellowpages; introduces progressive disclosure and routes the agent to the correct category router.

yp-auto-plan

Compatibility entrypoint for the old auto-plan workflow. Use `yp-workflow-plan` as the new default planning stage.

yp-brainstorm

Compatibility entrypoint for the old brainstorming workflow. Use `yp-workflow-frame` and `yp-workflow-design` as the new default path.

yp-tasks

Execution helper for task coordination in TASKS.md. Use from `yp-workflow-execute` when work benefits from explicit task claiming, status tracking, or worktree-aware coordination.

yp-verify

Compatibility entrypoint for the old verification stage. Use `yp-workflow-verify` as the new default verification stage.

yp-workflow-debugging

Use when software work stops being straightforward implementation and turns into investigation, root-cause analysis, or repeated failed fixes.

yp-workflow-design

Use when software work needs approach selection, scope boundaries, or a concrete design before planning or implementation.

yp-workflow-execute

Use when carrying out approved software work. Keeps execution minimal, surgical, and aligned with the approved plan or explicit user direction.

yp-workflow-frame

Use when starting software work to clarify the request, surface assumptions, define success criteria, and decide whether work should proceed to design, plan, or direct execution.

yp-workflow-git-worktrees

Use when software work should run in an isolated branch/worktree and later merge back cleanly.

yp-workflow-handoffs

Use when another agent or human needs a crisp summary of progress, blockers, verification, or next steps during software work.

yp-workflow-parallel-agents

Use when multiple independent software tasks or investigations can be delegated safely in parallel.

yp-workflow-plan

Use when the design or requirements are clear and the work needs to be decomposed into an execution plan with verification and task boundaries.

yp-workflow-review-loops

Use when software work benefits from lightweight quality gates between stages, tasks, or batches of changes.

yp-workflow-review

Use when software work needs correctness review, readiness checks, risk identification, or PR-oriented inspection after verification.

yp-workflow

Use for normal software-building workflow requests. Routes work into the six-stage lifecycle: frame, design, plan, execute, verify, review.

yp-workflow-subagents

Use when focused delegation to one or more subagents would improve software work without bloating the main session context.

yp-workflow-tdd

Use when test-first execution, regression-proof fixes, or stronger verification discipline would improve software work.

yp-workflow-verify

Use before any claim that work is complete, fixed, passing, ready, or safe to hand off. Requires fresh evidence before assertions.

convex-patterns

Conventions for writing Convex backend functions, designing schemas, handling auth and multi-tenancy, and managing errors. Use when: writing Convex queries, mutations, or actions, designing a Convex schema, adding indexes, setting up multi-tenant access…

frontend-architecture

Conventions for React app structure, TanStack Router file-based routing, state management separation (Zustand vs Convex vs TanStack Query), data fetching patterns, and component organization. Use when: building React pages, adding routes, choosing where to…

manage-global-skills

View and manage all globally installed skill libraries — inventory, install, remove, update.

manage-project-skills

View and manage the current project's skill context — .agents/, CLAUDE.md, TASKS.md, and installed skills.

monorepo-setup

Workspace structure, shared package conventions, build pipeline, and deployment configuration for Bun + Turborepo + Convex + Vercel monorepos. Use when: setting up a new project, configuring workspaces, adding a shared package, setting up build scripts,…

preferred-stack

Defines the preferred technology stack, architectural choices, and tooling for all new projects. Routes to domain-specific skills for backend, frontend, UI, and monorepo conventions. Use when: choosing a tech stack, starting a new project, making an…

react-patterns

React coding patterns and anti-patterns for this stack. Use this skill when writing React components, form state, data-sync hooks, or when reviewing code for useEffect misuse, form synchronization, or key-based remounting patterns.

scaffold-skill

Create a new yellowpages-compliant skill from a name and description.

ui-component-system

Design system conventions for the shared component library built on Radix UI headless primitives, Tailwind CSS v4, and class-variance-authority (CVA). Covers component patterns, semantic color tokens, dark mode, and the full component catalog. Use when:…

validate-skill

Run the yellowpages quality checklist against a skill and report pass/fail per criterion.

yp-compress

Rewrite a memory file in terse prose to reduce input tokens.

yp-context

Full transparency view — everything injected into this session at startup.

yp-diagnose

Scan all installed skills for yellowpages compliance violations and emit exact fix instructions.

yp-forget

Remove a note from CLAUDE.md Agent Notes section.

yp-help

Use when the user asks what yellowpages can do or how to find the right yellowpages skill.

yp-notes

Print current CLAUDE.md contents — make implicit agent memory explicit.

yp-reload

Re-read CLAUDE.md and installed skills state into conversation context via tool use.

yp-remember

Append a persistent note to CLAUDE.md Agent Notes section for future sessions.

yp-session

Model info, estimated context pressure, and active hooks for current session.