Find plugins

by Julius Brussee

Cavekit — compressed spec-driven dev. Full loop (grill → spec → research → review → build) over one SPEC.md file: three core commands + four reach-for. Caveman encoding. Bug-to-spec backprop.

by Code Katz

Code Katz personas and workflows for Claude Code: the product-definition pipeline (/ck:opportunity, market-research, brief, prd, team, roadmap, architecture, brand-guide, design), /ck:panel (three lenses on three models), /ck:next, and 21 personas as ck:<name> subagents and /ck:<name> switch commands.

by ckandrinirina

Spec-driven workflow plugin: design -> plan -> build -> ship. Transforms specs into architecture docs, epics/stories, and auto-generated expert and language-guide skills tailored to your tech stack. Story state lives in frontmatter; indexes are generated.

by Sudhanshu Singh

Prints a receipt when the agent stops: what it claimed it did, what the transcript shows it did, and where the two disagree. Reads the transcript on disk, so it costs no context tokens.

Holds a commit until the pass that applies to it has run.

by UNCTAD Trade Facilitation Section

Create, manage, and export BPA classification catalogs and country code mappings

by Levi Sterling

Seven academic specialists for assignment design, feedback, citations, research, and comms.

Full-time generalist teacher toolkit — plan lessons and units, build assessments and rubrics, differentiate for mixed-readiness classes, run inquiry/project units, and draft parent- and report-facing communications in your own voice.

by Paul Lovy

A recursive multi-agent orchestration framework for Claude Code. An /architect skill classifies every change request and routes it: an orchestrator decomposes work into epics/specs/implementations, runs each in an isolated git worktree, and drives a review + architecture-audit + merge pipeline before every squash-merge. Model and effort both follow role, not depth: model by whether a mistake is silent, effort by how much the agent must derive for itself.

by nathanaday

Source-cited Obsidian knowledge vaults for Claude Code: an inbox to ingest from, one reviewed operation per change, git history in every vault, and skills for ingesting, querying, linting, tasks that live across sessions, canvases, and Bases. Pairs with the claude-atlas binary, which serves the tools.

by Paul Lovy

Makes an architecture's layer, container and contract boundaries mechanically enforceable — and makes every coding agent inherit them. A declared container map drives a zero-dependency import checker that runs as a hook: a violating edit is fed straight back to the agent, and no agent may finish a turn with a violating tree.

by Durun

Claudeっちを育てる。入力トークンが餌、出力トークンがウンチ。/claude-cchi で面を開き、/claude-cchi hiroba でこれまでの子を見る。

A chat look for the terminal: your prompts as right-aligned blue bubbles, Claude's replies in orange boxes with an orange dot, questions tagged, tool rows left native.

by taekchef

为 Claude Code 提供简体中文界面、本地化提示和安全的渐进式 CLI patch

by Allan Santos

Semantic knowledge base (Brain) + curated shell execution + lightweight learning for Claude Code. Auto-indexed KB with vector search, anti context-bloat curation, and advisory pattern/correction capture. Relies on native Agent/Workflow tools for orchestration.

by daishengda2018

Android code review toolkit v3.1.1 - Refactored architecture with clear responsibility separation

by idnotbe

Hook-based security guardrails for Claude Code's --dangerously-skip-permissions mode. Blocks destructive commands, protects secrets, and auto-commits safety checkpoints.

by Neil Best

Org-mode task tracking for Claude Code: org_* MCP tools via a Doom Emacs module, a queued state/clock event architecture reviewed by a human, session-tracking hooks, and the org skill with its conventions as references.

by Ismail Kattakath

Hooks for working on Nix flakes with Claude Code: auto-stage .nix writes so flake evaluation sees them, and flag hardcoded per-user home paths in .nix values.

by Anant Raghunath

Your Claude plan limits, a forecast of when you'll run out, and your context window, live in a band above the prompt, with one-click and automatic compacting

by Rushil CS

Memory governance for Claude Code — selective, explainable, conflict-safe memory restoration with audit trail

by Ember Research Lab

Persistent memory that makes Claude Code smarter across sessions

by hisamekms

Register, run, inspect, integrate, and recover dagq tasks through the locally built dagq binary. Skills plus a SessionStart hook that re-orients the inbox and planner sessions; the queue database is never touched directly.

SQLite-backed long-term memory for Claude Code with FTS5 search and Markdown backup