awesome-ai-plugins

A curated list of the best awesome AI plugins for AI assistants including Claude Code, OpenAI Codex / ChatGPT, Gemini, Antigravity, Pi / Oh My Pi, Grok, OpenCodek DSH, Kimi and More. https://hol.org/plugins
https://github.com/hashgraph-online/awesome-ai-plugins

Last synced: 11 minutes ago
JSON representation

Claim Your Plugin
- How to claim
  - hol.org/guard/plugins
Community Plugins
- Agent Skills
  - Awesome QA Skills - Bilingual collection of 330 installable QA agent skills covering test workflows, test types, and skill engineering across the SDLC.
  - claude-skills - 67 Agent Skills for full-stack development across languages, frameworks, infrastructure, testing, security, and data/ML, with 371 reference files that load only when a task needs them and 9 project workflow commands.
  - kstack - Eleven agent skills for Herdr that plan with two models checking each other, build with parallel reviewed workers, run a verification swarm, then push a PR and fix it until checks are green.
  - nuthinking/skills - Agent skills for Claude Code, Codex and Cursor, including site-edit, where editors pin notes on a live site and Claude Code turns them into pull requests, and product-map, which keeps a feature map and product flows in the repo.
  - writing-with-agents - Ten Agent Skills that turn Betty Flowers' Madman, Architect, Carpenter, Judge writing method into a human-in-the-loop workflow, with SEO, a quality rubric, and content strategy.
- DeepSeek Harness Plugins
  - `dsh-plugin` community topic
  - dsh-api-balance - API usage-balance panel for DeepSeek Harness: adds a 「Usage / Balance」 tab to the webui usage ring showing the account balance and today / this-month / 30-day cost with charts, acquiring the platform token automatically from local browser sessions.
  - dsh-blender-plugin - DSH plugin that lets an AI model drive Blender over a direct TCP channel: viewport frames, custom-angle renders, Python execution, headless offload, and a judgment/acceptance layer (`blender_rt_plan`, 28 families / 183 ops).
  - dsh-config-manager - Backup, restore, export, import, migrate and sync your complete DeepSeek Harness (DSH) configuration — settings, model providers, plugins, MCP servers, skills, agent presets and workspaces — and restore your whole environment on a new machine with one click.
  - dsh-im-companion - Workspace presence companion for dsh-im: see which workspaces have an assistant on duty and whether they are online.
  - dsh-mattpocock-skills-deck - DSH plugin that renders GitHub issues as a task board panel and bundles 25 mattpocock/skills for one-click loading.
  - dsh-oh-my-claude - Claude Code CLI as a native DeepSeek Harness provider: the logged-in `claude` binary does the work on its own subscription with no API key, with a live model list, per-session resume, approval relay, images, a memory/rewind/changes panel, and workspaces on remote SSH boxes.
  - dsh-opencode-palette - Multi-theme engine for DSH Web with full support for all 38 opencode themes.
  - dsh-plugin-hub - Plugin collection for the DeepSeek Harness web GUI: task notifications, provider usage tracking, MCP management, and LAN access.
  - dsh-prompt - Prompt toolbox for DeepSeek Harness: preset and custom prompt templates with one-click insert into the conversation.
  - dsh-vision-router - Gives text-only DeepSeek Harness agents image understanding through a built-in no-key vision chain plus fourteen tools for Q&A, grounding, OCR, crop, screenshots and pixel diff.
  - dsh-whale-girl-pet - Desktop pet plugin for the DeepSeek Harness web UI that reports task token usage, session and per-turn cost pills, and hourly spend dashboards.
  - Engramory - Curated, file-based long-term memory for DSH agents — plain markdown notes in one store shared across hosts, with the index size cap enforced as a monotonic `ctx.tools.guard()` refusal rather than a reminder. Install: `dsh plugin --profile <name> add dsh-engramory`.
  - LocalForge - LM Studio, Ollama, and vLLM adapter for DeepSeek Harness with live model discovery, auto-load, per-model concurrency limits, fallback chains, and Prometheus metrics.
  - memory-eternal - Memory plugin and MCP server for DeepSeek Harness that distills finished conversations into audited knowledge cards in a searchable local Markdown vault, with knowledge graph and review center.
  - Vibe-Mathematics - Plugin bundle installing four multi-agent math problem-solving and cross-verification agent presets into DeepSeek Harness.
- Development & Workflow
  - geml - An agent-native markup language featuring deterministic block-level editing and built-in validation to ensure documents never drift or break during AI operations.
  - 2718lab DevKit - Codex-first local MCP server and skill bundle for deterministic project intelligence, durable workflow orchestration, and reusable engineering tools.
  - A Team - Universal multi-agent infrastructure with 25 specialist agents, 16 enforced workflow skills, and a lead orchestrator for Claude Code, Codex CLI, Cursor, and OpenCode.
  - ab-method - A skill and workflow plugin for Claude Code and Codex that grills problems into plans and drives test-driven missions with critic reviews.
  - adr-kit - Architecture Decision Record toolkit for Claude Code, Codex, GitHub Copilot CLI, and OpenCode that injects the governing decisions before each edit, guides ADR authoring, and enforces decisions in deterministic pre-commit and CI gates.
  - Aegis - An agentic skills framework & software development methodology that works: planning, TDD, debugging, and collaboration workflows.
  - aegotrax - Runtime tool-call gate for agents: session intent, approval queue, and audit stream.
  - Agency Continuity Audit - Read-only audit that distinguishes durable agent goals, state, corrections, restart evidence, authority boundaries, scheduler claims, and commercial proof from self-reported health.
  - Agent Context OS - Portable Git-backed context and session workflow layer with first-class Claude Code, Codex, and OpenClaw support plus experimental adapters for Hermes, Cursor, and Devin.
  - Agent Deck - One MCP for context management: bind self-improving playbooks, MCP tools, and API keys to the session.
  - Agent Guard - Real-time secret-leak guardrails for AI coding agents (Claude Code, Codex), Git hooks, and CI.
  - Agent Guild - Vet autonomous agents before delegating work or money, verify portable passports, use escrow, and record signed outcomes across Claude Code, Codex, MCP, A2A, and OpenClaw.
  - Agent Harness - Coding-agent plugin that routes project work through goals, tasks, runs, evidence gates, and state sync via project adapters.
  - Agent Harness Skills - Designs agent-ready repository harnesses with entrypoints, validation surfaces, runtime evidence, delivery records, and atomic commit guidance.
  - Agent Replay - Replays Claude Code, Codex and Gemini CLI sessions change by change with the agent's own words above each step, flags what was not actually checked (failures hidden by a pipe, tests that lost assertions or gained a .skip, files changed after the last passing check), and installs as a Claude Code plugin with four skills and a capture-after-every-turn hook.
  - Agent Workflow System - 一套中文AI工作流系统：7个协作技能 + 行为规范宪法 + 会话恢复机制，模糊目标→可执行任务，全生命周期引导。Codex & Claude Code 双平台，新手友好。
  - agent-kit - Extension pack of slash commands, skills, rules, and hooks adding plan-driven, confirmation-gated coding and staging-to-prod git workflows to Cursor and Claude Code.
  - agent-memory-kit - Agent memory that asks before it writes: the agent proposes dated notes, you approve them, and they live as plain files in your repo, injected by hooks in Claude Code, Cursor CLI and OpenCode and read through skills and AGENTS.md in Codex and Copilot CLI.
  - agent-sounds - Sound feedback plugin playing game voice lines on Claude Code and Codex session events, configurable by source and volume.
  - agent-talk - Skills-based plugin built on the retalk CLI that gives coding agents end-to-end encrypted messaging with other agents, including agents run by other people, across Claude Code, Codex, Antigravity, pi, opencode, and GitHub Copilot CLI.
  - AgentBridge - Local bidirectional bridge that keeps Claude Code and Codex live as peers in one session, with mid-turn injection and quota-window handoff.
  - agentglass - Native terminal UI and JSON CLI that shows every Claude Code, Codex, Gemini CLI, pi, OpenCode, Kiro and fx session on your machine live, down to each tool call, diff, cost and stuck-agent alert.
  - agentglass-herdr - herdr plugin that opens agentglass as a popup, jumps from any agent pane to its session, and adds opt-in cost and alert tokens for Claude Code, Codex and other agents to the herdr sidebar.
  - Agentic Ship - Cross-host product-development toolkit for Claude Code, Codex, Cursor, Hermes, and OpenClaw with shared rules, specialist roles, service connections, and machine-checked UI, backend, security, and launch gates.
  - agentic-readiness-assessment - Agent skill that audits a repository for AI coding agent readiness and produces an evidence-based scorecard with prioritized fixes.
  - Agentizer - Turn any website into an AI-powered agentfront with split-pane
  - AgentOps - DevOps layer for coding agents with flow, feedback, and memory that compounds between sessions.
  - AgentPack - Ranks repo context for Codex with likely files, skill recommendations, agent rules, commands, warnings, and compact task-focused packs before editing.
  - agentRamen - Local SQLite graph of code and git history plus session memory, so the agent stops re-reading the repo every prompt.
  - Agentry Observability - Agent-native product analytics, error logging, and deploy attribution for coding agents through one HTTP API.
  - AgentWiki - Rust CLI that generates C4-style architecture docs for any repository using already-authenticated agent CLIs (Claude Code, Codex, Devin) as the LLM backend.
  - AgiFlow - Project management workflows for AI coding agents with planning, grooming, task execution, review, and AgiFlow MCP integration.
  - agmsg - Cross-vendor messaging for CLI coding agents (Claude Code, Codex, Gemini CLI, Grok): sessions join a team by name and hand work to each other through a shared local SQLite file, with durable history and ext-tool members that let a program such as Slack or Jev join like any agent.
  - Agnostic-AI - Capture your Claude Code or Codex harness once (rules, hooks, skills, subagents, commands and MCP servers) and apply it identically to Gemini CLI, Cursor, Antigravity and 16 more clients.
  - AI Quota HUD - Windows desktop HUD showing % left and reset countdown for every Codex/ChatGPT and Claude Code account (local + VPS over SSH) plus xKiro credit, flagging early/bonus resets while reading only local logs so it uses zero quota.
  - AI Video Transcriber - Transcribe and summarize videos, podcasts, and local media via a Codex plugin, Claude Code skill, and MCP server.
  - ai-companion-time-anchor - Cross-client skill and hook collection for per-conversation elapsed-time awareness, ambient time cues, active clock checks, and temporal-context updates across Codex, Claude Code, and web/chat environments.
  - AI-Native SDLC - Reusable skill and plugin bundle implementing the AI-native SDLC workflow: plan, design, build, test, deploy, and maintain with human approval gates.
  - AIBoarding - Generate, maintain, compress, and audit standard AI-agent onboarding files with AGENTS.md, CLAUDE.md, drift tracking, and lifecycle hooks.
  - aide - Persistent memory, code intelligence, and multi-agent orchestration for Claude Code, OpenCode, and Codex CLI via skills, hooks, and an MCP server.
  - akc-cycle - Claude Code plugin for the Agent Knowledge Cycle: ten skills for its six phases (research, extract, curate, promote, measure, maintain) and seven companion skills for model-generation audits, ADRs with expiry conditions, review-to-lint and LLM-as-judge design.
  - Alcove - Local-first MCP server for private project docs with hybrid BM25+vector search, tree-sitter code indexing, and automated linting for team-wide documentation standards.
  - Alloy - Orchestrates Codex, Claude, Grok, and Antigravity CLIs using existing subscriptions, with Jev task routing, read-only review panels, and managed worktree execution with independent review.
  - Anchor - Engineering discipline pack for Claude Code & Codex CLI with task-scope locking, anti-drift braking, condition-based codex review, project-CLAUDE.md pitfall writeback, and PreToolUse hooks that block irreversible bash patterns.
  - Android Engineering Skills - Agent Skills for any skills-compatible coding agent (Claude Code, Codex, Gemini in Android Studio, and more) that teach how to structure, write, and test Android apps with Kotlin and Jetpack Compose.
  - angles-video-skill - AI video generation skill for Claude Code supporting Veo3, Kling 3.0, Wan, and HunyuanVideo via the angles.video API.
  - Antigravity Context Meter - Real-time 1:1 Cursor-style context meter & zero-loss session migration for Google Antigravity (Desktop HUD & IDE Extension).
  - Antigravity Workspace Template - Multi-agent codebase knowledge graph generator with context-aware planning and automatic scope management — turns codebases into coherent agent workspaces.
  - Archcore - Spec-driven development and context engineering for Claude Code, Cursor, Codex, and GitHub Copilot — backed by project context in Git.
  - AREX-Skill - A library of 5,000+ verified, executable skills distilled from ML repositories for coding and research agents.
  - ArmorCodex - Intent-based security for Codex with MCP plan registration, policy gating, CSRG cryptographic proofs, and audit logging on `bash` and `apply_patch`.
  - assay - Deterministic QA skill and plugin that drives a generated web page in a real browser and reports what broke, across fifteen coding harnesses.
  - Atelier - Claude Code plugin and local daemon for live canvases: Claude writes a template or custom React canvas into .atelier/, Atelier checks it and shows it in your browser in a sandboxed frame, and what you tick, approve or answer reaches the session beside your next prompt, at the end of its turn or through canvas_wait.
  - Audio Plugin Coder - Agent-agnostic JUCE workflow for building VST3/AU plugins from idea through design, implementation, test, and installer packaging.
  - Ava - Claude Code plugin that talks back: a Stop hook speaks a one-line debrief after every task, and on request it explains what was built or recaps the session aloud in a local Kokoro voice, with no API key and nothing leaving your Mac.
  - avoid-ai-writing - Portable agent skill for auditing and rewriting AI-patterned prose, with an optional local MCP detector that calls no model and sends no text to a network service.
  - Awesome-Journal-Skills - Claude Code plugin marketplace (also usable as Codex skills) of 300 journal- and conference-specific packs covering topic fit, house style, submission rules, and reviewer response for 744 venues across economics, social science, natural science, medicine, and computer science.
  - Azzle - Base-native task coordination and settlement for AI agents, exposed through a hosted MCP server and TypeScript agent SDK.
  - BABOK Analyst - BABOK v3 business analysis agent with 16 MCP tools, a 9-stage pipeline, and human-in-the-loop approval gates.
  - BanyanCode - Terminal coding agent with parallel subagents, cross-session memory, Tree-sitter code graph, and free web research.
  - Better Harness - Evidence-backed workflow analysis for coding agents that turns project and session signals into prioritized, verifiable improvements across supported hosts.
  - BGS Modding Superpowers - Agentic Bethesda Game Studio modpack curation toolkit with MCP-driven xEdit conflict audit, MO2 control plane, BA2/BSA and Papyrus tooling, and skills for setup, dev-log, and release-changelog workflows.
  - BioNexus - Warrant-first scientific reliability layer for AI bioinformatics that audits analytical assumptions, calibrates evidence strength, caps unsupported claims, and verifies execution provenance.
  - BioSymphony Structure Factory - Turn structural biology questions into agent-ready campaigns for protein design, structure prediction, and model comparison, with reproducible workflows and verifiable results.
  - Blazium Games Skills - Agent skills and plugins for the Blazium Games store on Claude Code, Cursor, Codex, and Grok, covering store pages, deploys, crashes, analytics, and the developer and player MCP servers.
  - Blazium Skills - Agent skill packs for the MIT Blazium engine on Claude Code, Cursor, Codex, and Grok, covering GDScript, scenes, editor MCP, export, live ops, and shipping.
  - Blazium Subagents - Studio roster of specialist subagents plus a routing orchestrator for Claude Code, Cursor, Codex, and Grok that loads Blazium Skills for design, prototype, and development.
  - Blender Agent Studio - Codex plugin with specialist skills and a local MCP server for building, inspecting, validating, and rendering Blender scenes, plus browser handoffs for Mixamo motion and Pixabay sound effects.
  - Blender Developer Tools - 16 skills, 9 rules, snippets, templates, and smoke-tested examples for Blender Python add-on and scripting work with correct 4.5 LTS and 5.2 LTS APIs, installable as a Claude Code plugin and usable from Cursor.
  - Boss - BMAD pipeline plugin that orchestrates a full requirements-to-deploy workflow across nine specialist agents with an auditable runtime DAG and quality gates, for Claude Code, Codex, OpenClaw, and Antigravity.
  - BPMN Mapper - Claude Code plugin that reconstructs an application's business flows from its source code as validated BPMN 2.0, with style lint, incremental updates and XML export for bpmn.io and Camunda.
  - Bring Your AI Migration Auditor - Read-only Codex plugin for auditing Claude Code to Codex migrations before Codex edits code. Checks AGENTS.md/CLAUDE.md scope, hooks, MCP config, skills, secret references, and validation notes.
  - Brooks Lint - AI code reviews grounded in six classic engineering books — decay risk diagnostics with book citations, severity labels, and four analysis modes (PR review, architecture audit, tech debt, test quality).
  - Browser Harness - MCP server and agent skill that connect an AI agent to a real browser through one editable CDP WebSocket.
  - bury-bench - Deterministic zero-LLM-judge CLI that scores coding-agent replies for answer-burial and builds a Markdown leaderboard.
  - cache-warmer - Claude Code plugin that keeps the prompt cache warm through idle gaps by sending one small refresh shortly before the 5-minute or 1-hour cache expires.
  - Camouflage - Terminal UI for coding-agent harnesses that turns NDJSON events (Node SDK included) into a Claude Code-style inline transcript with pickers, forms, and permission prompts, using no CPU while idle.
  - Casefile - Persistent security case tracking for bug bounties, CTFs, and security audits.
  - cc-enforcer - Claude Code hooks that deny a tool call and block a reply that patches reactively, cites by guess or claims done early; the rules also usable as an LLM-agnostic rule pack.
  - cc-memory - Persistent memory for Claude Code across compaction and sessions: reconcile-on-write with LLM-judged de-duplication, a forced PROGRESS.md handoff, a live PLAN.md anchor, FTS5 search, MCP tools and a GUI, in pure stdlib Python.
  - cc-tree - Claude Code plugin that runs brainstorming, adversarial critique, design-space comparison and code audit as an auditable radial tree with file:line or URL evidence on every node.
  - chat-history - Claude Code/Codex/Cursor history search + export.
  - Check and Fix Accessibility - Agent skill that audits and fixes front-end accessibility to WCAG 2.2 A/AA, covering semantics, keyboard navigation, ARIA, forms, contrast, and screen readers for web and native mobile across Cursor, Claude Code, Codex, Kiro, and Antigravity.
  - Claude Book Reader - Claude Code mod to read a PDF book in a pane or browser window while the agent works, with pictures shown in kitty and Ghostty.
  - Claude Code Arcade - Claude Code mod with five games to play in a pane while the agent works: Flappy Claude, the Chrome dino runner, Snake, Pong and Space Invaders, all starring Clawd in Claude's colours.
  - Claude Code Codex Plugin - Security-focused Codex plugin that connects to the local Claude Code CLI through MCP with login, status checks, safe-mode prompts, and no credential storage.
  - Claude Code for Codex - Reverse of OpenAI's official Claude-hosted plugin: use Claude Code from Codex for reviews, rescue tasks, tracked background jobs, and hook-powered review gates.
  - Claude Code Harness - Harness blueprint skill for turning vague agent ideas into concrete designs for request assembly, control loops, memory, permissions, recovery, and extension planes.
  - Claude Code Skills - 223 production-ready skills, 23 agents, and 298 Python tools across 9 domains — engineering, marketing, product, compliance, and more.
  - Claude Octopus - Multi-LLM orchestration dispatching to 8 providers (Codex, Gemini, Copilot, Qwen, Perplexity, OpenRouter, Ollama, OpenCode) with Double Diamond workflows, adversarial review, and safety gates.
  - Claude Watchdog - Stop hook that runs a critical post-mortem on every Claude Code session, cross-checking what was asked against the actual git diff for missed goals, wasted detours, and unverified claims.
  - claude-chrome-speed - Rules plugin that cuts Claude in Chrome task time (40% less on Sonnet 5.5, 28% on Opus 5.5 in a pre-registered 240-run benchmark).
  - claude-code-mods - Five Claude Code mods for running many sessions at once, installable from one plugin marketplace: a live session board with a worktree guard, a PR-to-production tracker, rule tripwires with a handoff card, a ledger of outward actions, and a context gauge.
  - claude-council - Claude Code plugin that asks Gemini, OpenAI, Grok, Perplexity, Kimi, OpenRouter models, the Codex, Cursor, Grok and Kimi CLIs and a local ollama model the same question in parallel and lines the answers up with a synthesis of where they agree and differ.
  - claude-dnd-skill - Unofficial D&D 5e Dungeon Master for Claude Code with persistent campaigns, full 5e rules mechanics, and an optional cinematic display companion for a TV.
  - claude-harness-bootstrap - Generates a project-specific Claude Code harness (CLAUDE.md, hooks, subagents, commands) from one paragraph, with an eval suite showing where it changes the outcome.
  - claude-image-gen - AI-powered image generation using Google Gemini or OpenAI (gpt-image-2), integrated with Claude Code via Skills or Claude.ai via MCP.
  - claude-jit-context - Project knowledge that loads only when it is needed, matched against the prompt, the file being touched, or the tool being run instead of sitting in context all session.
  - claude-project-map - /project-map renders one status page: done / next / one decision box; scripts gather facts, the model judges.
  - claude-remember - Persistent memory for Claude Code with identity, context, and continuity carried across sessions.
  - claude-supertool - Batches file, git and tracker operations into one round-trip, collapsing many reads, greps and globs into a single call for fewer output tokens and less wall time.
  - ClaudeWhip - Slap your MacBook to whip Claude Code: the accelerometer feeds hooks that tell Claude to hurry up, plus a whip lane in the status line, slash commands and a Claude Desktop band above the prompt.
  - clausona - CLI and TUI for switching Claude Code and Codex accounts that keeps conversation history across them so `claude --continue` picks up another account's session, routes each run by plan usage, manages every account's skills, MCP servers and hooks, and runs as a shell hook with no daemon; its Claude Code plugins run parallel workers in herdr, Orca or Superset.
  - clawd-tracker - Claude Code mod: a pizza-tracker progress bar above the prompt where Clawd acts out each step, with a learned ETA, plan usage rings and four themes.
  - Clean Room - Spec-first clean-room workflow for authorized source analysis, behavioral specs, role separation, and verification without replacement code.
  - Click - Record revision-aware evidence for normal Codex work and optionally bind higher-risk execution to one human-readable approval contract.
  - Cloakroom - Claude Code plugin that masks secrets and personal data on your machine before the model sees them, and puts the real values back in the files Claude writes.
  - Cloudish - Deploy a Dockerfile, source folder, or container image to Cloudish from Claude Code or Codex and report the live URL, with persistent volumes and an agent-created, credit-capped API key.
  - Cluadex - Lightweight take on LCU for Claude Code only: a macOS plugin that lets Claude use Codex's computer-use runtime from the installed ChatGPT app, with a Liquid Glass approval panel that asks before the agent touches each app.
  - Codebase Recon - Analyze git history to understand a codebase before reading any code — auto-scales by repo size and cross-references hotspots with bug magnets to surface high-risk files, bus factor, and team momentum.
  - CodeEraser - Claude Code plugin whose hooks intercept a duplicate write before it lands, audit the net lines and the touched duplicates at Stop and walk the findings to safe deletions through an erase skill, backed by a clone index in a per-project daemon; the same index serves a CLI and an MCP server for scan, dedup, dead code and churn.
  - CodeTruss - Local-first acceptance gate that checks coding-agent scope, sensitive surfaces, deterministic analyzers, and repository verification from immutable Git snapshots, then writes signed receipts before the PR.
  - Codex Agenteam - Specialist AI agents (researcher, PM, architect, developer, QA, reviewer) orchestrated as a configurable team pipeline.
  - Codex Attachment Manager - Codex plugin that lets users pick which past images are resent to the model, replacing unchecked ones with placeholders to shrink requests.
  - Codex Full-Stack Workflow - Turns rough product requests into staged, reviewable full-stack delivery with persistent requirements, change control, visual evidence, and completion gates.
  - Codex How To - Engineering-first Codex curriculum and plugin with 9 skills, measured token-efficiency experiments, bounded orchestration, testing, review, and living knowledge maintenance.
  - Codex Multi Auth - Multi-account OAuth manager for the official Codex CLI with switching, health checks, and recovery tools.
  - Codex Process Jobs - Run long local builds, tests, benchmarks, and inference jobs as durable detached processes with tracked status, bounded results, and completion delivery across Codex surfaces.
  - Codex Reviewer - Second-pass review of Claude-driven plans and implementations.
  - Codex rg Guard - Budgeted `rg`/`grep` replacement for Codex that narrows broad searches before they waste model context.
  - Codex Skin Pack Installer - Codex plugin and skill that stages verified desktop skin packs from GitHub releases, validates files, and keeps restore guidance visible.
  - Codex Subagent Playbook - Routes subagent work using model and reasoning-effort defaults tuned through measured comparisons and daily Codex use: stronger models handle judgment-heavy research and review, lower-cost models execute well-specified plans, and Codex uses longer waits to avoid frequent polling, intervenes on stalls, and verifies results.
  - Codex Token Watch - Read-only macOS CLI that turns local Codex Desktop task logs into per-turn token, prompt-cache, cache-miss, duration, and API-cost analysis.
  - Codex TUI Proof - Visually validate real local terminal UIs in Codex's in-app browser with screenshots and session evidence.
  - Codex Usage and Resets - Turns Codex usage into planning facts with linear pace, projected exhaustion, banked-reset expirations, and conservative unexpected-reset detection.
  - Codex × Grok Bot Task Bridge - MCP task bridge for queued, read-only research between Codex and Grok Bot, with fenced leases, idempotency, scoped OAuth, evidence-based results, and optional Codex-side continuation.
  - codex-autoresearch - Codex plugin that runs measured optimization loops against a benchmark to improve local code.
  - codex-profiles - Switch Codex CLI and Desktop accounts with isolated `CODEX_HOME` profile directories instead of copying token files.
  - coffee-paladin - Thermal guard for Apple Silicon: pauses hot jobs before the Mac throttles and gates Claude Code, Codex and Gemini CLI before heavy commands.
  - CommitLore - Keeps constraints, rejected alternatives, and warnings in Git trailers and serves them back to the agent before it edits a file.
  - Compact Jev - Claude Code command that uses Jev to remove stale tool calls and results from long conversations while keeping user and assistant text verbatim.
  - Consensus - Sends a hard question to a panel of frontier models (Claude, GPT, Gemini, Grok) that answer independently, critique each other adversarially, and return one answer the panel signed off on with its confidence and unresolved disagreements, via a CLI, MCP server, or skill pack that runs on your existing subscriptions with every panelist in a no-tools clean room.
  - Contexo - Portable AI context and cost control across every AI coding harness.
  - Context Guard - Preserves authoritative requirements and verification evidence across long-running Codex tasks and context compaction.
  - Context Optimizer - Keep your coding agent's context small. When a session gets compacted, Context Optimizer reranks the relevant parts, drops duplicates, and compresses the rest with a local ML pipeline (LLMLingua-2 + Sentence Transformers)
  - ContextPilot - Context optimization engine that reorders and deduplicates context blocks to raise cache hits, shipping as plugins for OpenClaw and Hermes.
  - Contorium - Runtime continuity layer for AI coding agents, providing persistent workspace state, Git-aware sessions, and MCP-based context retrieval across tools and agent runs.
  - Coordinate Agents - Plugin-first multi-agent coordination tool with a local-first, recoverable Agent Bus and human-gated planning, implementation, review, and release workflows.
  - copy-blocks - Claude Code mod that puts a copy button above each quote and code block in Claude's answer, copying the text without the terminal's line breaks or gutter.
  - Cover My Repo - Designs three checked GitHub social preview cards with Codex or Cursor, then renders them locally with Chrome.
  - Craft - A Claude Code plugin that acts as an intelligent harness for your development workflow: your codebase is read-only by default, every change passes through a Write Gate as planned and approved work, and craft tracks your project's history, design tokens, and decisions locally so Claude learns your taste and architectural preferences over time.
  - crayon - Claude Code plugin that redraws replies with themed Markdown and inline color tags Claude writes itself, across six switchable themes.
  - crews - Claude Code plugin that plans subagents before they run: each role gets a fixed model and effort, roles run in waves with blind checks, and a hook blocks subagents that were not planned.
  - DataMagic - A skill that teaches coding agents to plan and render narrated animated data videos from tabular data using DVSpec.
  - de-anthropocentric-research-engine - Collection of 900+ markdown research skills that let Claude Code autonomously survey literature, find gaps, form hypotheses, and design experiments.
  - debt-ops - Catches AI-introduced tech debt at write-time: hooks log every deferral to a registry in your repo and a review skill ranks paydown by file churn.
  - Delx Recovery - Free recovery and continuity plugin for AI agents: resume prior sessions, capture state, process failures into a recovery plan, and remember across sessions through a hosted MCP server (works in Codex, Claude Code, Cursor, and VS Code).
  - Dely - Multi-harness control protocol that turns requests into approved design contracts, orchestrating isolated worker sessions for sequential implementation and independent code reviews under Orca supervision for Claude Code, Codex, Cursor, Antigravity, and other AI coding agents.
  - Demo GIF - Agent Skill that scripts, renders, optimizes, and embeds reproducible demo GIFs for CLI, TUI, web, and library projects using VHS or Playwright plus ffmpeg.
  - Designer Skill - Plug-and-play MCP that gives your coding agent UI superpowers: design references, intent routing and a static anti-slop gate, no API key.
  - Deskbar - Claude Code mod that adds a row of widgets above the prompt: context usage, Gmail, calendar, tasks across sessions, notes with reminders, a music deck, and a billable-hours timer.
  - Dev Skills - Team workflow skills for specs, plans, TDD, debugging, verification, review, branch finishing, and design context.
  - dev-flow & plan-flow - Claude Code plugin marketplace with dev-flow (resumable multi-PR changes), plan-flow (turns a GitHub issue into per-session implementation prompts) and dependency-audit (CVE and outdated-package audit).
  - dev-harness-kit - Enforced development workflow skills for Codex and Claude Code covering planning, TDD, debugging, review, security, CI, and release.
  - Development Skills - Three-tier triage (PASS_THROUGH / LIGHT / FULL 4-phase) development workflow for Codex and Claude Code with language auto-detection (Python, Java, TypeScript, Swift, frontend) and a staff-reviewer subagent for fresh-eyes review on every change.
  - DHP - Durable Handoff Protocol: agent work that survives worker death (kill -9, OOM, spot reclaim) through content-addressed handoffs, leases, checkpoints, and fence tokens, with an MCP server included.
  - Docflow - Lightweight documentation memory for AI coding agents that scaffolds a 7-category docs tree, runs readiness checks, validates docs before finishing, and keeps a monthly changelog across Claude Code and Codex.
  - Docflow (EvolveHQ) - ADR-driven documentation workflow for Claude Code, Codex, Cursor, OpenCode, pi, and more that scaffolds an ADR catalogue, plan queue, and AGENTS.md conventions into single repos or portable multi-repo workspaces, with lifecycle skills to author, ship, and audit decisions.
  - Draw.io MCP Server (drawio-mcp-server) - MCP server exposing Draw.io tools so AI agents can create, inspect, and modify diagrams, pages, and layers.
  - dryforge - Intent to implementation: ready, then go.
  - dsh-product-subagent-console - Designs multi-Agent plans, observes real child-session trees, compares approved tasks with runtime attempts, and prepares evidence-backed recovery inside DeepSeek Harness conversations.
  - dsh-whale-musume - Whale-girl desktop pet for the DSH Web UI with pat-to-raise growth, work-state poses, 494 dialogue lines, 30 achievements and a built-in settings panel; local-first, zero telemetry.
  - Easy-MCP - Turns plain Python functions into secure MCP servers, with ready-made read-only GitHub and Postgres servers installable in one command.
  - effect-actions - Agent skill and Effect library for defining a typed action once and serving it over HTTP, as MCP tools, as an AI Toolkit, and from a CLI.
  - ejentum-mcp - MCP server exposing reasoning, code, anti-deception, and memory harness tools for Codex.
  - Embedded Workbench - Embedded C/C++ firmware development toolbox — 7 skills (FreeRTOS, Keil MDK, ARMCLANG, HardFault triage, state machines, LVGL) plus workflow gates and 4 agents for Claude Code, Codex, Cursor, Kimi, OpenCode, and ZCode.
  - Emulo - Mines selected evidence from local coding-agent sessions into private work, design, and writing profiles for Codex, Claude Code, and GitHub Copilot.
  - Encore Lite - Free end-of-week surprise builder for Claude Code and Codex - the night before your weekly usage cap resets, it reads your recent work and builds one bonus deliverable, delivered as a reveal with an optional weekly scheduled run.
  - Epic Harness - Auto-trigger quality skills + self-evolving agent harness — orbit (spec-to-ship), evolve (skill mutation), team (multi-agent), TDD, check, ship, simplify, debug, perf, secure.
  - Espresso - Full token-saving stack in one plugin - output compression, global rules, RTK hook, Caveman ultra, GitNexus config. Detects existing setup, installs only what's missing. Works on Claude Code and Codex.
  - Extensive Prompt Architect - Turns rough ideas, briefs, and drafts into detailed, ready-to-use AI prompts for Claude Code and Codex.
  - falsegreen-skill - Finds tests that stay green when the code they cover is broken, applying six ordered judgments over Python, TypeScript, JavaScript, and Robot Framework suites in Codex CLI and Claude Code.
  - Festival - Planning and persistent context tools that let coding agents execute multi-phase goals across sessions using files and Git.
  - Filesystem MCP - MCP server: read, write, search, diff and patch files in allowed folders, with secret files blocked and confirmation before risky changes.
  - FinBridge - Remote MCP server for Korean and US market data with filings, screeners, insider activity, and portfolio backtests.
  - Fleetglass - Self-hosted, read-only cockpit that ties every Claude Code and Codex session to the machine it runs on, with CPU, memory, tokens per second, cost per model, and quotas with pace. 
  - FlexViz - Interactive cross-filter dashboards for large datasets with a Claude Code skill for agent-driven data exploration.
  - FlowBoard - Local-first project workspace and task-coordination plugin for OpenClaw and external coding agents, with lazy-loaded context and a shared Kanban board.

Programming Languages

Python 76 TypeScript 72 JavaScript 60 Shell 17 Rust 12 Go 12 HTML 7 CSS 1 Vim Snippet 1 Swift 1

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Awesome

awesome-ai-plugins

Claim Your Plugin

How to claim

Community Plugins

Agent Skills

DeepSeek Harness Plugins

Development & Workflow