Development plugins

Scan a codebase, review it against a chapter-backed checklist, size the next 10x with a deterministic calculator, and write design documents with trade-offs; 28 condensed case studies.

6 skills

Detects the CMS a project publishes to, writes a machine-readable cms-publish.json spec describing its content types, fields, auth, and transport, then publishes content through that spec. Works with file-based systems (Astro, Hugo, Eleventy, Jekyll), headless APIs (Sanity, Contentful, Payload, Strapi, Directus), and hosted CMSes (WordPress, Ghost). Also humanizes AI-sounding drafts before they go out.

6 skills

開発全般で使える汎用スキル集。Git/GitHub ワークフロー、テスト設計、DDD のレイヤ分割、GitHub Actions の検査、textlint の ai-words 辞書への語の登録をカバー

6 skills

Multi-agent UX review system with simulated user personas and specialist analysis. Orchestrates intake interviews, persona research, app walkthroughs, and technical/visual specialist reviews into a prioritized backlog.

1 skill · 1 command · 4 agents

Terminal session management - fork sessions, run commands in panes/tabs, capture output

1 skill · 4 commands · hooks

Consistent code delivery from AI agents. Test-driven development, git worktrees, merge-conflict resolution, receiving code review and lint pipelines, plus a code-reviewer agent.

5 skills · 1 agent

Hooks-based tracer that logs prompts, tool calls, tool results, and assistant responses.

5 commands · hooks

Feature development agents for design, implementation planning, and scaffolding

3 commands · 3 agents

Two-tier prompt orchestration: Haiku for discovery/refinement, Sonnet/Opus for execution. Uncovers real problems through Socratic questioning, refines context, and optimizes prompts before expensive model execution. Saves 60-80% on clarification costs.

2 commands · 4 agents

Mobile-web game planning, player flows, HUDs, game feel, optional Figma UI, FSD/ECS Three.js greybox prototypes, lint-gated pixel sprite animation, lifelike cat behavior with a trace check, and three.js pixel-art dioramas with a draw-call budget.

6 skills

Full punch plugin (docx report pipeline + corpus research, completed from eplus-punch-reports v0.4.0) with the vo modifications under test: corpus lookups delegated to a budgeted Sonnet 5 punch-researcher subagent, and UI-visible hook diagnostics. No bundled MCP/token; needs a managed punch-query server.

3 skills · 1 command · 1 agent · hooks

Go backend development automation for Venturo skeleton - Guided workflows for features, entities, endpoints, and adapters

6 commands

Type-agnostic multi-agent review toolkit. Spawn N clean-context reviewers to evaluate any artifact against requirements. Iterative review-fix cycles with deterministic convergence.

3 skills · 3 agents

Engineering practice skills for Grok. TDD, dual-suite SDD, code review, check-work verification, and document-code-contracts.

6 skills

Claude account switching, quota status and rotation policy via ccbroker: /ccb-status, /ccb-use, /ccb-auto and /ccb-policy slash commands plus a SessionStart hook that runs 'ccb pull' (requires the ccb binary on PATH).

5 commands · hooks

GitHub workflow commands for issues, PRs, and project management using the gh CLI

6 commands

Build full-stack Databricks Apps with React + FastAPI. Provides MCP tools for dev server management, component search, SDK docs, and deployment.

4 skills · 1 MCP server · hooks

Fail-closed agentic SDLC governance: issue-first lifecycle, path budgets, one approval from another account, and a governed merge.

6 skills

Hook di sicurezza, skill commerciali e review agent per progetti LivenBit

4 skills · 1 agent · hooks

Work one GitHub issue: a planner (Opus) first reads it and the code and posts an implementation plan or the open owner questions on the issue, or closes the issue when the owner's answer picks a close option, then the worker implements the current plan to an open pull request, or unsupervised through independent review and merge; an orchestrator runs planner and worker for labelled or named issues and continues issues that are already started (current plan, own pushed branch, own open PR) until they are merged; a skill proposes the opt-in label per issue and sets it on confirmation.

4 skills · 2 agents

Cut LLM API costs: prompt caching audits, fingerprinting, cost estimates and batch planning. Companion to the LLM Token Engineering blog post.

6 skills

Build AI product screens with Tattva's React parts: find, plan, build and review through the Tattva MCP server.

1 skill · 4 commands · 1 MCP server

Connects your agent to the Glomo MCP server: search the docs, find and inspect API operations, generate sample requests, get the ordered calls for an integration, and run calls against your sandbox account.

5 skills · 1 MCP server

Build, seed, test and improve Cavelon solutions from a repository: the Cavelon skills and the cavelon MCP server.

4 skills · 1 MCP server · hooks

LSP-backed code intelligence for Claude Code via MCP and native LSP - navigate definitions, references, call hierarchies, type hierarchies, and diagnostics across codebases

2 skills · 1 agent · 1 MCP server · hooks · LSP

Spec-driven development on OpenSpec: fresh-agent spec review, a light conformance review of code against its spec (with a mutation spot check), fixes routed through the spec, mechanical archive PRs, /sdd:init to set a repo up for the flow, and /sdd:opt-out to mark one that won't use it

4 skills · 2 commands

Deliberate continuity between contexts: /handoff (/hh) writes a four-field resume note in docs/handoffs/, kept out of git; /pickup (/hp) resumes the newest after /clear or in a new session and deletes it, so notes never merge into main; in cloud sessions the note can be committed and pushed so it outlives the container, and /handoff:relay (/hr) continues the work in a new cloud session instead of a /clear. Adapted from Jordan Loyd's keystone (MIT)

6 commands

Sally, the lifecycle manager for every skill in this repo.

6 skills

Analisa o projeto, verifica dependências, diagnostica e configura a máquina do usuário, recomenda ajustes e instala tudo para rodar o projeto com gate obrigatório de confirmação.

1 skill · 4 commands · 1 agent

Playbooks, named principles, multi-model review, and proof on the real app.

3 skills · 2 agents · hooks