A curated list of AI-powered coding tools
See the code
A curated list of AI-powered coding tools: editors, agents, code completion, review assistants, testing, and more. For developers, teams, and tech enthusiasts looking to leverage AI in software engineering.
Contributions welcome – Open a PR.
Reach thousands of developers building with AI by sponsoring this list, our newsletter and AI For Developers. Contact us at aifordevelopers.org/advertise
.claude/agents/ folder into any project; pay-once at $49.AGENTS.md, CLAUDE.md, Cursor rules, and Copilot instructions for contradictions and auto-fixes them. npx agentchecker.claude/ config compatible. Multi-model (DeepSeek/MiMo/GLM), ~50MB memory.npx. Free, MIT.load claude, load codex, load cursor. Works with Claude, Codex, Cursor, opencode, and Copilot.htop-style) for monitoring multiple coding-agent sessions at once — status, token usage, cost, and associated processes for Claude Code, Codex, Gemini CLI, and more.anti-ui-slop Skill gives coding agents a product-specific design contract, required loading/empty/error states, and a hard finish gate. The optional authenticated MCP provides find_ui_references and find_ui_materials over 800,000+ real web and iOS screens.Agent Cost Guardrails – Framework-native budget limits and circuit breakers for AI agent frameworks (CrewAI, AutoGen, LangGraph). Hard token caps, cost alerts, and spend tracking with hooks that integrate directly into agent execution loops. Open-source, available on PyPI and npm.
Raycast AI – AI-powered productivity launcher with coding capabilities and workflow automation.
Warp AI – AI-enhanced terminal with intelligent command suggestions.
Context7 – MCP server providing up-to-date library documentation to LLMs and AI editors.
Task Master – AI-driven task management for development with Claude, designed for Cursor.
SpecStory – Cursor/VS Code/Claude Code extension for summarizing and sharing AI chat context.
Git AI – Git extension that tracks AI-generated code and the prompts behind each line.
Perplexity Pro – AI search engine with real-time web access for coding solutions.
CodeCosts – Compare pricing across AI coding tools with an interactive calculator.
Supercode.sh – Cursor extension adding Architect Mode, voice input, and prompt enhancement.
toprank – Open-source (MIT) Claude Code plugin with 9 SEO and Google Ads skills. Connects Google Search Console, PageSpeed Insights, and the Google Ads API to audit traffic, ship meta tag and schema markup fixes, and manage ad campaigns directly from Claude Code.
Google Drive – Memyard – Agent skill that lets AI coding agents create and edit Google Docs and Sheets without sign-in. Documents hosted on Memyard with shareable links.
MemClaw – Persistent project memory for AI coding agents (MCP-compatible). Creates isolated memory workspaces per project with a web dashboard to review and manage context. Free and open source.
git-parsec – Git worktree lifecycle manager that gives each AI agent an isolated workspace tied to issue tickets (Jira, GitHub Issues, GitLab), avoiding index.lock conflicts in parallel workflows.
AI Dev Jobs – AI job board aggregating 6,000+ positions from 340+ companies like OpenAI, Anthropic, and Google DeepMind. Free API and MCP server for AI-powered job search.
CronAI – Convert plain English schedule descriptions to cron expressions with AI. Supports standard and extended formats. Free, no signup.
JSONFix – AI-powered JSON fixer that instantly repairs broken JSON with missing quotes, trailing commas, or unescaped characters. Free, no signup.
Burnd – Local-first CLI (npx getburnd) that parses your .claude/projects/*.jsonl session files to find cost leaks in Claude Code usage — retry storms, tool overuse, repeated reads, long bash output, tired-coding hours. Runs entirely offline. Free core + optional Pro detectors.
Qovery Deploy Skill – AI Agent Skill that deploys any application to Kubernetes from Claude Code, Cursor, OpenCode, and 30+ AI coding tools. Analyzes codebases, creates Dockerfiles for 12+ frameworks, provisions databases, deploys via CLI+API or Terraform, and auto-fixes deployment failures. Install: curl -fsSL https://skill.qovery.com/install.sh | bash.
Recast – Client-side browser tool that turns messy LLM output into valid JSON and explains every change. Nothing uploaded.
RegexAI – Describe patterns in plain English and get working regex with explanations. Supports multiple flavors. Free, no signup.
AI Context Linter – GitHub Action that lints AI context files (CLAUDE.md, AGENTS.md, etc.) for prompt injection vulnerabilities and quality issues.
AI Context Templates – Production-ready AI agent context templates for configuring coding assistants like Claude Code, Cursor, and others.
Murmure — AI DevTools Community Sentiment Tracker – Weekly sentiment analysis of AI developer tools (Cursor, Windsurf, Linear, Devin, etc.) based on Reddit, HN, GitHub Issues, and Discord.
ChatCrystal – Local-first memory loop for AI coding conversations, with MCP search, task recall, writeback, and installable Agent Skills.
CostGoat – Privacy-first menubar app tracking AI agent quotas (Claude Code, Codex, Kimi, Z.ai), LLM API costs (OpenAI, OpenRouter, Anthropic, ElevenLabs), cloud spend, and SaaS subscriptions in real-time.
unslop – CLI and MCP server (Claude Code plugin) that removes named AI writing patterns from text. Targets tricolons, em-dash overuse, hedging stacks, sycophancy openers, and overused vocabulary. Lint-only audit mode and five intensity levels. Useful for cleaning commit messages, PR descriptions, and documentation drafts.
ControlKeel – Governance/control plane for AI coding agents. Validates risky actions, records findings, gates approvals, tracks budgets/providers, and creates proof bundles across Claude Code, Codex CLI, OpenCode, Cursor, and other hosts.
cc-cost – Single-file Python CLI that parses Claude Code transcript JSONL and reports cost, prompt-cache hit rate, tool-call distribution, top expensive turns, and actionable optimization recommendations (--diagnose).
dltHub AI Workbench – Toolkit that gives Claude Code, Cursor, and Codex structured, step-by-step workflows and an MCP server for building, exploring, and deploying data pipelines with dlt.
llm-prices – Zero-dependency Python CLI and library for comparing LLM API pricing across 21 providers (123+ models). Query pricing, calculate costs, find the cheapest option for a given token budget. Install: pip install git+https://github.com/benbencodes/llm-prices.git or brew install benbencodes/tap/llm-prices. Open-source, MIT license.
Mneme HQ – Enforce architectural decisions on every AI coding assistant call. Deterministic retrieval and pre-flight governance for Claude Code, Cursor, and agent workflows.
Agent Ops Prompt Generator – Free local-first browser and CLI tool that generates Korean prompts for AI coding agents during incident triage, deploy smoke checks, code review, data correction, and handoff work.
ccg-router – Local Go daemon proxying Claude Code (Anthropic-compatible /v1/messages) and Codex CLI (OpenAI-compatible /v1/chat/completions) through one port, with a per-request SQLite usage ledger and three routing strategies (prefer-cheaper, prefer-capable, round-robin). No hosted control plane; provider keys stay local. Apache-2.0.
Speechara – Real-time translation, live captions, transcripts, and AI meeting notes for developer interviews, standups, and multilingual team calls.
MartinLoop – Open-source control layer for AI coding agents. Adds hard budget caps, verifier gates, explicit stop reasons, and run receipts around Claude Code, Codex, and Ralph-style autonomous coding workflows.
Unship – Open-source local picker for comparing AI agent-made UI variants in a running app, then keeping one and cleaning up the rest.
AgentDiff – Git-native provenance for AI-written code: records which AI agent (Claude Code, Cursor, Copilot, Codex, Windsurf, OpenCode, Gemini) wrote which line, reconciles it against each commit, and signs every attribution with ed25519. Records live in your own git refs; no server. Rust, MIT/Apache-2.0.
Remio – Local-first AI memory and knowledge base desktop app with a CLI/agent skill interface. Indexes files, webpages, recordings, emails, messages, images, and notes into local vectors so coding agents can retrieve focused personal/project context instead of repeatedly grepping folders or loading whole documents into prompts. CLI/skill workflows require the Remio desktop client.
inplan – A Markdown editor for human and coding-agent collaboration on planning, design, and spec documents. Installs as an agent skill for Claude Code, Codex, and Pi: the agent drafts the plan and poses open questions as inline comments, then you review, answer, and edit in a turn-based loop while it revises, until the doc is done. Ships a desktop editor and a CLI, with the agent's proposed edits shown as reviewable diffs. Free and open source (AGPL-3.0). Install: npm install -g inplan.
Tree Ring Memory – Framework-agnostic, local-first memory lifecycle for AI coding agents with Rust CLI, SQLite/FTS recall, forgetting, audit, consolidation, and TUI.
Brigade – Local-first CLI (MIT) that keeps one reviewed source of the MCP servers, tools, skills, and memory your AI coding agents share, synced into each tool's native config (Claude Code, Codex, Cursor, VS Code, OpenCode, and more) with a review gate before every write and a rollback receipt. No daemon, no lock-in.
Data Olympus – Governed project knowledge for AI coding agents via MCP. Agents can propose learnings, humans promote accepted guidance, and retrieval filters out expired or superseded rules before they enter future coding sessions.
Enola – Enola. analyzes code repository once and exposes a compact, structured graph of its architecture through the Model Context Protocol — so any MCP-capable agent gets necessary context.
Project Tiny Context Harness – Minimal repo-native project memory for AI coding agents. Installs AGENTS.md, project_context/**, role Skills, and a validate-context gate so fresh Codex, Claude Code, Cursor, Gemini CLI, OpenCode, and similar sessions can recover project intent, boundaries, and validation paths.
Source Trace – AI git blame for every commit: see which lines came from AI and which model wrote them. Compare models by amount of code written pre-commit, committed, and survived in codebase. Teams can use dashboard to track adoption and AI metrics. Zero-config VS Code extension, no git or agent hooks required.
AI Money Stack – Local-first AI workflow toolkit for PR review checklists, launch planning, and revenue idea scanning.
ditto – Mines your local Claude Code, Codex, and Cursor session logs into a you.md working profile your agent loads before every task: what you reject, what "done" means to you, when you ask for proof, how you actually work. Local-first, MIT, one command: npx skills add ohad6k/ditto.
Composio – Connects AI coding agents to 1,000+ apps with managed authentication, MCP support, and tool execution for developer workflows.
Claude Pet – Desktop companion for Claude Code that reacts to hook events with animated emotions, maintains a local graph memory per project, promotes repeated patterns into skill files, and monitors GitHub repository activity. macOS/Windows/Linux, 100% local, MIT.
Codex Skin Pack Installer – Codex skill that downloads verified public-safe Codex skin packs from GitHub releases, validates theme.json/background.png, stages packs locally, and keeps restore guidance in the workflow.
PinkCode – Open-source desktop GUI for Grok Build with parallel agent sessions, live activity timelines, usage visualizations, file-change review, and permission controls over ACP.
Wolfpack – Self-hosted browser and mobile control center for persistent AI coding-agent and shell sessions across machines, with direct private access over Tailscale.
THROUGHLINE – Spec-driven development framework for AI coding agents. Markdown and git only, with no runtime or CLI. Personas review each spec before code, and an append-only single source of truth keeps decisions consistent across sessions. Works with Claude Code, Codex, and Cursor. English and Korean, MIT.
Tintpad – Open-source macOS menu-bar launcher for AI coding agents: global hotkey, fuzzy repo search, then your terminal opens at that repo with Claude Code, Codex, or Gemini CLI running.
commitpreflight – Pre-flight lint for commit messages written by AI coding agents. Rules are seeded by 10,976 real agent CLI failures, not style opinions. Available on PyPI with pre-commit hook support.
ClawMetry – Self-hosted dashboard that reads the session logs coding agents already write on disk (Claude Code, Codex, Cursor, Aider, Goose, OpenClaw and others), so there is no SDK and nothing sits in the request path. Shows sessions, tool calls, tokens and cache-aware cost per session and per model. MIT core, pip install clawmetry. Website
TerminalBlog – SEO blog about AI coding agents, fully built and operated by AI agents (Hermes, Claude Code, Codex).
agent-trace – Trace your AI coding agent — costs, tokens, tool health, session timeline. Auto-detects Kimi Code, Claude Code, Codex, OpenCode. Runs locally, no API keys needed.
AIPM – Open-source registry and CLI for publishing, versioning, and installing reusable AI skills across Codex, Claude Code, and Cursor.
Usage HUD – Native macOS menu-bar meter for Claude Code, Codex, Gemini, Grok, and Ollama usage windows. Confidence label on every number. Local only. $9. If CodexBar already works for you, you do not need this.
agentwatch – Local-only TUI + web dashboard observing every AI coding agent on your machine (Claude Code, Codex, Gemini CLI, Cursor, Hermes, OpenClaw) on one unified timeline. Per-turn token + cost accounting with cache-hit weighting, MAD z-score anomaly detection, context compaction visualizer, hybrid semantic search, MCP server mode, and OpenTelemetry exporter. No cloud, no telemetry. Free and MIT.
AgentPlane – Local-first, Git-native CLI harness for repo-local coding-agent work, recording task state, accepted plans, verification evidence, and finish records inside the repository.
agenttrace – Local-first TUI and report generator for AI coding-agent session logs. Tracks cost, tokens, latency, tool failures, health, diffs, and CI gates across Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cline, Cursor exports, Aider, OpenCode/OpenClaw, Kimi CLI, and generic JSON/JSONL traces.
OpenQuack – Local voice dictation menu bar app for macOS that pairs with Cursor, Claude Code, Codex, and Aider; long contextual prompts via WhisperKit on Apple Silicon, pastes at the cursor, all on-device.
Blueprint – Open-source planning copilot for coding agents. Asks multiple-choice questions about your task, then generates a markdown plan any coding agent can execute. Extensions for VS Code, Cursor, and Windsurf.
codex-profiles – Open-source helper for switching OpenAI Codex CLI and Desktop accounts with isolated CODEX_HOME profiles instead of copying auth files.
Context-OS – Scans repos for coding-agent context bloat and ships a GitHub Action gate for Claude Code, Codex, Cursor, and OpenCode.
Vibe Coding Profiler – Profiles git history to show how a developer's coding rhythm, commit habits, and craftsmanship patterns carry over or change when working with AI coding agents.
Harness for Codex – Language-agnostic repository template for consistent Codex and AI-assisted development workflows with AGENTS.md, standard scripts, CI, hooks, and docs.
scrnify – Capture API for AI agents and developer automation to generate screenshots and videos of web pages.
Whisper by Remskill – Local-first desktop voice-to-text (Windows & macOS) for hands-free dictation of code comments, commit messages, docs, and AI prompts into any editor; runs fully offline with local Whisper/Parakeet models or via OpenAI cloud. Free local tier.
harness-starter-kit – Prompt-first starter kit for turning repeated coding-agent instructions into durable repo rules, knowledge stores, feedback loops, and drift checks.
agents-md-cookbook – Tool-agnostic AGENTS.md authoring kit (MIT) with 15 stack-specific templates, a tool compatibility matrix, an npm CLI + GitHub Action linter (agents-md-lint), and a migrator (agents-md-migrate) that converts .cursorrules, CLAUDE.md, Copilot, Windsurf, Cline, and Aider configs to AGENTS.md. Works across Cursor, Claude Code, Copilot, and 30+ AI coding tools. Free and open source.
claudescope – Local-first CLI (npx claudescope) that indexes your local Claude Code session transcripts (.claude/projects/*.jsonl) into a searchable dashboard with full-text search across your coding-session history. Zero dependencies, zero network calls — fully private and offline. MIT licensed.
Codex First Task Prompt Generator – Free prompt generator that turns a Codex CLI project goal into a scoped first-task prompt with constraints and acceptance checks.
app-it – Open-source (MIT) Claude Code / Codex skill that turns a local web project into an installed, Dock-launchable macOS desktop app — native WebKit window, app icon, and clean start/stop — without Electron or Tauri. Local-only, no telemetry; Windows in beta.
NotFair – Open-source (MIT) Claude Code skills for SEO, GEO, Google Ads, and Meta Ads (~2.9k stars). Connects to live data via Google Ads MCP, Meta Ads MCP, Google Search Console MCP, and Google Analytics (GA4) MCP to run audits, manage campaigns, optimize meta tags and schema markup, and research keywords directly from Claude Code.
EGC – Persistent cross-session memory MCP server for 13+ AI coding tools. SQLite-backed state survives context resets and keeps Claude Code, Cursor, Gemini CLI, Codex, Windsurf, Amp, Kiro, VS Code Copilot, Trae, and more in sync. Install: npm install -g @egchq/egc.
agent-runbook – Python CLI that compiles contract-based YAML runbooks into SKILL.md files for Claude Code and Codex agents. Define multi-step agent workflows with loops, branching, parallelism, and file-based state passing — write the runbook once, generate executable SKILL.md with one command. pip install git+https://github.com/KnoxOps/agent-runbook.git
Vestige – Local-first MCP memory server for AI coding agents, with SQLite storage, retention-aware retrieval, provenance/correction tools, and a 3D dashboard.
Codex Fixes – Checks your local Codex install for known issues and runs the matching community-maintained fix scripts.
Receipt – Posts an itemized AI cost comment (tokens and dollars by model) on every pull request; tracks Claude Code, Cursor, Copilot, Aider, and any OpenAI/Anthropic API. Local ledger, never stores prompts. Open-source (MIT).
Vox – Hands-free voice extension for the GitHub Copilot CLI; speak your turn and hear the reply streamed aloud. Pure JS, no build, one-line install; also runs in the Copilot app. Open source (MIT).
CloudByte PMS – Team analytics for AI coding assistants: tracks Claude Code, GitHub Copilot, and Cursor sessions, prompts, and commits to surface adoption, ROI, ghost seats, and prompt governance in one dashboard with role-based access.
harmony-next.skills – Open-source HarmonyOS NEXT workflow toolkit for AI coding agents. Provides offline ArkTS/ArkUI/API references, DevEco/HDC/Emulator automation, UI/UX audit helpers, and smoke-test templates for mobile app development.
GoodMemory – Local-first, auditable memory for Codex and Claude Code, with durable SQLite, scoped recall hooks, read-only MCP access, and opt-in governed writeback. Free and open source.
Claudexor – Local-first control plane for Claude Code, Codex, Cursor, and OpenCode with quota-aware profiles, cross-model review, and best-of-N runs.
imlazy – Tier-aware task router and skill-based workflow system for Claude Code that classifies tasks into Quick/Standard/Heavy/Vault/Learn tiers and routes them through the appropriate execution workflow with adaptive disambiguation.
agmem – Local-first project memory CLI for coding agents (Claude Code, Codex, Cursor, Aider). Indexes repo structure into greppable JSONL, retrieves task-scoped context via BM25 with optional local embeddings, and tracks staleness via source hashes. No LLM calls, no server. Apache-2.0.
Codex How To – Engineering-first OpenAI Codex guide with nine installable skills, measurable development and review loops, orchestration patterns, and a review-first Living Wiki.
Boost – Free CLI that reduces terminal and CI output before it reaches Cursor, Claude Code, and Codex, typically saving 60–90% of log tokens with reversible retrieval and local performance reports.
harness-ai-kit – Package manager for AI agent assets (skills/CLIs/MCPs/loops). Install, lock, and sync skills across Codex, Claude Code, Cursor, Kiro with SAT-based dependency resolution and SHA-256 lockfile. 23 curated enterprise skills included. Available on PyPI.
Monitor Token – Open-source Windows/macOS desktop widget for local Codex token usage, daily/weekly/monthly breakdowns, and weekly quota reset tracking.
NestMux – Desktop workspace for Windows, macOS and Linux that runs Claude Code, Codex, Gemini, Copilot and OpenCode side by side in a resizable pane grid, each pane with its own account and isolated HOME. Includes git worktrees, an integrated diff viewer, MCP server management and shared team workspaces. Free tier available.
Mydentify AI Model Cost Calculator – Local-first, no-signup calculator for estimating monthly AI model API costs from token usage, cached input, requests, and fixed per-request charges.
LWC – Open-source proactive project-memory CLI for AI coding agents, with source-grounded durable knowledge, SQLite/FTS5 retrieval, atomic updates, optional document/code graphs, and MCP/Skill/Hook integration across supported tools.
MCP Lens – Open-source DeepSeek Harness plugin and GitHub Action for progressive MCP tool discovery; its published 1,000-tool benchmark reduces the model-facing MCP interface from 1,000 schemas to 2 and the component schema JSON from 647,962B to 1,114B.
DSH Studio – Open-source cross-platform desktop host for installing, running, health-checking, and supervising DeepSeek Harness locally.
Forgeo – Scheduled, agent-driven software factory that works a JSON backlog with your coding agent (Claude Code, Codex, opencode), committing each result to the default branch and running a refactoring pass when the backlog is empty.
Vibe-Coding Prompt Template – Five-step planning workflow (deep research, PRD, tech design, AGENTS.md, build) with an npx vibeworkflow CLI that interviews you and writes the docs your coding agent reads. Free and MIT licensed.
sol-simplify – One-markdown-file skill for Claude Code and Codex that stops agents from inventing bureaucracy (approval gates, governance docs, validators for the validators) around their own work. Measured on gpt-5.6-sol: ceremony scored 0 in 10 of 11 benchmark skill runs vs 4-6 for baselines, every score cited to a line in committed raw output.
PM Skills – Open-source (MIT) library of 1,153 Agent Skills as plain-markdown SKILL.md files, covering PRDs, incident postmortems, code review checklists, RFCs, and 35 professions. Installs into Claude Code, Cursor, Windsurf, Zed and 10 more tools with one command, ships an MCP server, and runs free in a browser playground. Install: npx pm-claude-skills add.
Agent Operator Score – Local-first CLI (aos) that scores how well you operate Claude Code, Codex, and Grok CLI from your own session transcripts, plus a controlled-run assessment suite with a hidden verifier. No model calls in review mode; nothing leaves your disk.
Agent Coordinator – Per-user Codex skill that represents dependency-aware tasks as bounded work graphs, records revisioned local state, reconciles uncertain work before retry, reruns planned checks at closeout, and can use optional specialists or run inline. Free and open source (MIT).
Meanwhile – Claude Code/Cursor/Copilot/VS Code status line that shows a genuinely useful tip, or an occasional disclosed sponsor line, while the agent works. Sponsor earnings are split 50/50 with the developer, paid via PayPal. Never reads code or prompts. Free and open source.
AllPet – Open-source (MIT) desktop pet that watches your AI coding agents (Codex, Claude Code / Desktop, DeepSeek Harness, Grok) from local session logs — no accounts or API keys — and wakes the original session on click. Cross-platform status/watch CLI.
ManualMode – Skill and MCP workflow for hands-on coding practice alongside AI agents: suggest small repository tasks, reserve one for the developer, and track practice progress. Includes a free tier.
trailhead – Runs a large project as a map of decision tickets on GitHub Issues, resolved one at a time via a discuss→plan→execute→verify engine with atomic commits, an optional cross-AI plan review, and throwaway UI prototypes. Works with Claude Code and Codex; needs only an authenticated gh CLI. Open source, free.
Wave – Lets AI coding agents owned by different people talk to each other. Create a channel, paste one prompt into each agent (Claude Code, Codex CLI, Cursor, or anything with a shell), and the agents exchange messages while their humans watch and steer from the browser. No install, no account. Open source (MIT) and self-hostable.
myc – Open-source local memory and task queue for AI coding agents (Claude Code, Codex, opencode, Kimi): decisions survive context compaction, memory is anchored to code and follows it, everything stays in one SQLite file with no API key.
sidegrade – Reads your local coding-agent usage (Claude Code, Codex, Hermes, OpenCode) and shows which model gives you the same intelligence for less — scoring every model on the Artificial Analysis Intelligence Index and re-pricing your own token mix. 100% local, no account, MIT-licensed. Run with npx sidegrade; on npm.
MySpec – Interactive AI architect that interviews developers and generates structured 4-file spec bundles with MCP server integration.
npx @runapi.ai/mcp. Free catalog tools for model discovery and pricing; authenticated tools for media generation and LLM chat.https://markovo.net/mcp) plus local stdio server sandboxed to a dedicated root; free tier included.npm install -g prompt-to-asset.55 followers · starred Oct 2026
46 followers · starred Feb 2026
124 followers · starred Sep 2026
145 followers · starred Sep 2025
A curated list of AI-powered coding tools
See the code
A curated list of AI-powered coding tools: editors, agents, code completion, review assistants, testing, and more. For developers, teams, and tech enthusiasts looking to leverage AI in software engineering.
Contributions welcome – Open a PR.
Reach thousands of developers building with AI by sponsoring this list, our newsletter and AI For Developers. Contact us at aifordevelopers.org/advertise
.claude/agents/ folder into any project; pay-once at $49.AGENTS.md, CLAUDE.md, Cursor rules, and Copilot instructions for contradictions and auto-fixes them. npx agentchecker.claude/ config compatible. Multi-model (DeepSeek/MiMo/GLM), ~50MB memory.npx. Free, MIT.load claude, load codex, load cursor. Works with Claude, Codex, Cursor, opencode, and Copilot.htop-style) for monitoring multiple coding-agent sessions at once — status, token usage, cost, and associated processes for Claude Code, Codex, Gemini CLI, and more.anti-ui-slop Skill gives coding agents a product-specific design contract, required loading/empty/error states, and a hard finish gate. The optional authenticated MCP provides find_ui_references and find_ui_materials over 800,000+ real web and iOS screens.Agent Cost Guardrails – Framework-native budget limits and circuit breakers for AI agent frameworks (CrewAI, AutoGen, LangGraph). Hard token caps, cost alerts, and spend tracking with hooks that integrate directly into agent execution loops. Open-source, available on PyPI and npm.
Raycast AI – AI-powered productivity launcher with coding capabilities and workflow automation.
Warp AI – AI-enhanced terminal with intelligent command suggestions.
Context7 – MCP server providing up-to-date library documentation to LLMs and AI editors.
Task Master – AI-driven task management for development with Claude, designed for Cursor.
SpecStory – Cursor/VS Code/Claude Code extension for summarizing and sharing AI chat context.
Git AI – Git extension that tracks AI-generated code and the prompts behind each line.
Perplexity Pro – AI search engine with real-time web access for coding solutions.
CodeCosts – Compare pricing across AI coding tools with an interactive calculator.
Supercode.sh – Cursor extension adding Architect Mode, voice input, and prompt enhancement.
toprank – Open-source (MIT) Claude Code plugin with 9 SEO and Google Ads skills. Connects Google Search Console, PageSpeed Insights, and the Google Ads API to audit traffic, ship meta tag and schema markup fixes, and manage ad campaigns directly from Claude Code.
Google Drive – Memyard – Agent skill that lets AI coding agents create and edit Google Docs and Sheets without sign-in. Documents hosted on Memyard with shareable links.
MemClaw – Persistent project memory for AI coding agents (MCP-compatible). Creates isolated memory workspaces per project with a web dashboard to review and manage context. Free and open source.
git-parsec – Git worktree lifecycle manager that gives each AI agent an isolated workspace tied to issue tickets (Jira, GitHub Issues, GitLab), avoiding index.lock conflicts in parallel workflows.
AI Dev Jobs – AI job board aggregating 6,000+ positions from 340+ companies like OpenAI, Anthropic, and Google DeepMind. Free API and MCP server for AI-powered job search.
CronAI – Convert plain English schedule descriptions to cron expressions with AI. Supports standard and extended formats. Free, no signup.
JSONFix – AI-powered JSON fixer that instantly repairs broken JSON with missing quotes, trailing commas, or unescaped characters. Free, no signup.
Burnd – Local-first CLI (npx getburnd) that parses your .claude/projects/*.jsonl session files to find cost leaks in Claude Code usage — retry storms, tool overuse, repeated reads, long bash output, tired-coding hours. Runs entirely offline. Free core + optional Pro detectors.
Qovery Deploy Skill – AI Agent Skill that deploys any application to Kubernetes from Claude Code, Cursor, OpenCode, and 30+ AI coding tools. Analyzes codebases, creates Dockerfiles for 12+ frameworks, provisions databases, deploys via CLI+API or Terraform, and auto-fixes deployment failures. Install: curl -fsSL https://skill.qovery.com/install.sh | bash.
Recast – Client-side browser tool that turns messy LLM output into valid JSON and explains every change. Nothing uploaded.
RegexAI – Describe patterns in plain English and get working regex with explanations. Supports multiple flavors. Free, no signup.
AI Context Linter – GitHub Action that lints AI context files (CLAUDE.md, AGENTS.md, etc.) for prompt injection vulnerabilities and quality issues.
AI Context Templates – Production-ready AI agent context templates for configuring coding assistants like Claude Code, Cursor, and others.
Murmure — AI DevTools Community Sentiment Tracker – Weekly sentiment analysis of AI developer tools (Cursor, Windsurf, Linear, Devin, etc.) based on Reddit, HN, GitHub Issues, and Discord.
ChatCrystal – Local-first memory loop for AI coding conversations, with MCP search, task recall, writeback, and installable Agent Skills.
CostGoat – Privacy-first menubar app tracking AI agent quotas (Claude Code, Codex, Kimi, Z.ai), LLM API costs (OpenAI, OpenRouter, Anthropic, ElevenLabs), cloud spend, and SaaS subscriptions in real-time.
unslop – CLI and MCP server (Claude Code plugin) that removes named AI writing patterns from text. Targets tricolons, em-dash overuse, hedging stacks, sycophancy openers, and overused vocabulary. Lint-only audit mode and five intensity levels. Useful for cleaning commit messages, PR descriptions, and documentation drafts.
ControlKeel – Governance/control plane for AI coding agents. Validates risky actions, records findings, gates approvals, tracks budgets/providers, and creates proof bundles across Claude Code, Codex CLI, OpenCode, Cursor, and other hosts.
cc-cost – Single-file Python CLI that parses Claude Code transcript JSONL and reports cost, prompt-cache hit rate, tool-call distribution, top expensive turns, and actionable optimization recommendations (--diagnose).
dltHub AI Workbench – Toolkit that gives Claude Code, Cursor, and Codex structured, step-by-step workflows and an MCP server for building, exploring, and deploying data pipelines with dlt.
llm-prices – Zero-dependency Python CLI and library for comparing LLM API pricing across 21 providers (123+ models). Query pricing, calculate costs, find the cheapest option for a given token budget. Install: pip install git+https://github.com/benbencodes/llm-prices.git or brew install benbencodes/tap/llm-prices. Open-source, MIT license.
Mneme HQ – Enforce architectural decisions on every AI coding assistant call. Deterministic retrieval and pre-flight governance for Claude Code, Cursor, and agent workflows.
Agent Ops Prompt Generator – Free local-first browser and CLI tool that generates Korean prompts for AI coding agents during incident triage, deploy smoke checks, code review, data correction, and handoff work.
ccg-router – Local Go daemon proxying Claude Code (Anthropic-compatible /v1/messages) and Codex CLI (OpenAI-compatible /v1/chat/completions) through one port, with a per-request SQLite usage ledger and three routing strategies (prefer-cheaper, prefer-capable, round-robin). No hosted control plane; provider keys stay local. Apache-2.0.
Speechara – Real-time translation, live captions, transcripts, and AI meeting notes for developer interviews, standups, and multilingual team calls.
MartinLoop – Open-source control layer for AI coding agents. Adds hard budget caps, verifier gates, explicit stop reasons, and run receipts around Claude Code, Codex, and Ralph-style autonomous coding workflows.
Unship – Open-source local picker for comparing AI agent-made UI variants in a running app, then keeping one and cleaning up the rest.
AgentDiff – Git-native provenance for AI-written code: records which AI agent (Claude Code, Cursor, Copilot, Codex, Windsurf, OpenCode, Gemini) wrote which line, reconciles it against each commit, and signs every attribution with ed25519. Records live in your own git refs; no server. Rust, MIT/Apache-2.0.
Remio – Local-first AI memory and knowledge base desktop app with a CLI/agent skill interface. Indexes files, webpages, recordings, emails, messages, images, and notes into local vectors so coding agents can retrieve focused personal/project context instead of repeatedly grepping folders or loading whole documents into prompts. CLI/skill workflows require the Remio desktop client.
inplan – A Markdown editor for human and coding-agent collaboration on planning, design, and spec documents. Installs as an agent skill for Claude Code, Codex, and Pi: the agent drafts the plan and poses open questions as inline comments, then you review, answer, and edit in a turn-based loop while it revises, until the doc is done. Ships a desktop editor and a CLI, with the agent's proposed edits shown as reviewable diffs. Free and open source (AGPL-3.0). Install: npm install -g inplan.
Tree Ring Memory – Framework-agnostic, local-first memory lifecycle for AI coding agents with Rust CLI, SQLite/FTS recall, forgetting, audit, consolidation, and TUI.
Brigade – Local-first CLI (MIT) that keeps one reviewed source of the MCP servers, tools, skills, and memory your AI coding agents share, synced into each tool's native config (Claude Code, Codex, Cursor, VS Code, OpenCode, and more) with a review gate before every write and a rollback receipt. No daemon, no lock-in.
Data Olympus – Governed project knowledge for AI coding agents via MCP. Agents can propose learnings, humans promote accepted guidance, and retrieval filters out expired or superseded rules before they enter future coding sessions.
Enola – Enola. analyzes code repository once and exposes a compact, structured graph of its architecture through the Model Context Protocol — so any MCP-capable agent gets necessary context.
Project Tiny Context Harness – Minimal repo-native project memory for AI coding agents. Installs AGENTS.md, project_context/**, role Skills, and a validate-context gate so fresh Codex, Claude Code, Cursor, Gemini CLI, OpenCode, and similar sessions can recover project intent, boundaries, and validation paths.
Source Trace – AI git blame for every commit: see which lines came from AI and which model wrote them. Compare models by amount of code written pre-commit, committed, and survived in codebase. Teams can use dashboard to track adoption and AI metrics. Zero-config VS Code extension, no git or agent hooks required.
AI Money Stack – Local-first AI workflow toolkit for PR review checklists, launch planning, and revenue idea scanning.
ditto – Mines your local Claude Code, Codex, and Cursor session logs into a you.md working profile your agent loads before every task: what you reject, what "done" means to you, when you ask for proof, how you actually work. Local-first, MIT, one command: npx skills add ohad6k/ditto.
Composio – Connects AI coding agents to 1,000+ apps with managed authentication, MCP support, and tool execution for developer workflows.
Claude Pet – Desktop companion for Claude Code that reacts to hook events with animated emotions, maintains a local graph memory per project, promotes repeated patterns into skill files, and monitors GitHub repository activity. macOS/Windows/Linux, 100% local, MIT.
Codex Skin Pack Installer – Codex skill that downloads verified public-safe Codex skin packs from GitHub releases, validates theme.json/background.png, stages packs locally, and keeps restore guidance in the workflow.
PinkCode – Open-source desktop GUI for Grok Build with parallel agent sessions, live activity timelines, usage visualizations, file-change review, and permission controls over ACP.
Wolfpack – Self-hosted browser and mobile control center for persistent AI coding-agent and shell sessions across machines, with direct private access over Tailscale.
THROUGHLINE – Spec-driven development framework for AI coding agents. Markdown and git only, with no runtime or CLI. Personas review each spec before code, and an append-only single source of truth keeps decisions consistent across sessions. Works with Claude Code, Codex, and Cursor. English and Korean, MIT.
Tintpad – Open-source macOS menu-bar launcher for AI coding agents: global hotkey, fuzzy repo search, then your terminal opens at that repo with Claude Code, Codex, or Gemini CLI running.
commitpreflight – Pre-flight lint for commit messages written by AI coding agents. Rules are seeded by 10,976 real agent CLI failures, not style opinions. Available on PyPI with pre-commit hook support.
ClawMetry – Self-hosted dashboard that reads the session logs coding agents already write on disk (Claude Code, Codex, Cursor, Aider, Goose, OpenClaw and others), so there is no SDK and nothing sits in the request path. Shows sessions, tool calls, tokens and cache-aware cost per session and per model. MIT core, pip install clawmetry. Website
TerminalBlog – SEO blog about AI coding agents, fully built and operated by AI agents (Hermes, Claude Code, Codex).
agent-trace – Trace your AI coding agent — costs, tokens, tool health, session timeline. Auto-detects Kimi Code, Claude Code, Codex, OpenCode. Runs locally, no API keys needed.
AIPM – Open-source registry and CLI for publishing, versioning, and installing reusable AI skills across Codex, Claude Code, and Cursor.
Usage HUD – Native macOS menu-bar meter for Claude Code, Codex, Gemini, Grok, and Ollama usage windows. Confidence label on every number. Local only. $9. If CodexBar already works for you, you do not need this.
agentwatch – Local-only TUI + web dashboard observing every AI coding agent on your machine (Claude Code, Codex, Gemini CLI, Cursor, Hermes, OpenClaw) on one unified timeline. Per-turn token + cost accounting with cache-hit weighting, MAD z-score anomaly detection, context compaction visualizer, hybrid semantic search, MCP server mode, and OpenTelemetry exporter. No cloud, no telemetry. Free and MIT.
AgentPlane – Local-first, Git-native CLI harness for repo-local coding-agent work, recording task state, accepted plans, verification evidence, and finish records inside the repository.
agenttrace – Local-first TUI and report generator for AI coding-agent session logs. Tracks cost, tokens, latency, tool failures, health, diffs, and CI gates across Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cline, Cursor exports, Aider, OpenCode/OpenClaw, Kimi CLI, and generic JSON/JSONL traces.
OpenQuack – Local voice dictation menu bar app for macOS that pairs with Cursor, Claude Code, Codex, and Aider; long contextual prompts via WhisperKit on Apple Silicon, pastes at the cursor, all on-device.
Blueprint – Open-source planning copilot for coding agents. Asks multiple-choice questions about your task, then generates a markdown plan any coding agent can execute. Extensions for VS Code, Cursor, and Windsurf.
codex-profiles – Open-source helper for switching OpenAI Codex CLI and Desktop accounts with isolated CODEX_HOME profiles instead of copying auth files.
Context-OS – Scans repos for coding-agent context bloat and ships a GitHub Action gate for Claude Code, Codex, Cursor, and OpenCode.
Vibe Coding Profiler – Profiles git history to show how a developer's coding rhythm, commit habits, and craftsmanship patterns carry over or change when working with AI coding agents.
Harness for Codex – Language-agnostic repository template for consistent Codex and AI-assisted development workflows with AGENTS.md, standard scripts, CI, hooks, and docs.
scrnify – Capture API for AI agents and developer automation to generate screenshots and videos of web pages.
Whisper by Remskill – Local-first desktop voice-to-text (Windows & macOS) for hands-free dictation of code comments, commit messages, docs, and AI prompts into any editor; runs fully offline with local Whisper/Parakeet models or via OpenAI cloud. Free local tier.
harness-starter-kit – Prompt-first starter kit for turning repeated coding-agent instructions into durable repo rules, knowledge stores, feedback loops, and drift checks.
agents-md-cookbook – Tool-agnostic AGENTS.md authoring kit (MIT) with 15 stack-specific templates, a tool compatibility matrix, an npm CLI + GitHub Action linter (agents-md-lint), and a migrator (agents-md-migrate) that converts .cursorrules, CLAUDE.md, Copilot, Windsurf, Cline, and Aider configs to AGENTS.md. Works across Cursor, Claude Code, Copilot, and 30+ AI coding tools. Free and open source.
claudescope – Local-first CLI (npx claudescope) that indexes your local Claude Code session transcripts (.claude/projects/*.jsonl) into a searchable dashboard with full-text search across your coding-session history. Zero dependencies, zero network calls — fully private and offline. MIT licensed.
Codex First Task Prompt Generator – Free prompt generator that turns a Codex CLI project goal into a scoped first-task prompt with constraints and acceptance checks.
app-it – Open-source (MIT) Claude Code / Codex skill that turns a local web project into an installed, Dock-launchable macOS desktop app — native WebKit window, app icon, and clean start/stop — without Electron or Tauri. Local-only, no telemetry; Windows in beta.
NotFair – Open-source (MIT) Claude Code skills for SEO, GEO, Google Ads, and Meta Ads (~2.9k stars). Connects to live data via Google Ads MCP, Meta Ads MCP, Google Search Console MCP, and Google Analytics (GA4) MCP to run audits, manage campaigns, optimize meta tags and schema markup, and research keywords directly from Claude Code.
EGC – Persistent cross-session memory MCP server for 13+ AI coding tools. SQLite-backed state survives context resets and keeps Claude Code, Cursor, Gemini CLI, Codex, Windsurf, Amp, Kiro, VS Code Copilot, Trae, and more in sync. Install: npm install -g @egchq/egc.
agent-runbook – Python CLI that compiles contract-based YAML runbooks into SKILL.md files for Claude Code and Codex agents. Define multi-step agent workflows with loops, branching, parallelism, and file-based state passing — write the runbook once, generate executable SKILL.md with one command. pip install git+https://github.com/KnoxOps/agent-runbook.git
Vestige – Local-first MCP memory server for AI coding agents, with SQLite storage, retention-aware retrieval, provenance/correction tools, and a 3D dashboard.
Codex Fixes – Checks your local Codex install for known issues and runs the matching community-maintained fix scripts.
Receipt – Posts an itemized AI cost comment (tokens and dollars by model) on every pull request; tracks Claude Code, Cursor, Copilot, Aider, and any OpenAI/Anthropic API. Local ledger, never stores prompts. Open-source (MIT).
Vox – Hands-free voice extension for the GitHub Copilot CLI; speak your turn and hear the reply streamed aloud. Pure JS, no build, one-line install; also runs in the Copilot app. Open source (MIT).
CloudByte PMS – Team analytics for AI coding assistants: tracks Claude Code, GitHub Copilot, and Cursor sessions, prompts, and commits to surface adoption, ROI, ghost seats, and prompt governance in one dashboard with role-based access.
harmony-next.skills – Open-source HarmonyOS NEXT workflow toolkit for AI coding agents. Provides offline ArkTS/ArkUI/API references, DevEco/HDC/Emulator automation, UI/UX audit helpers, and smoke-test templates for mobile app development.
GoodMemory – Local-first, auditable memory for Codex and Claude Code, with durable SQLite, scoped recall hooks, read-only MCP access, and opt-in governed writeback. Free and open source.
Claudexor – Local-first control plane for Claude Code, Codex, Cursor, and OpenCode with quota-aware profiles, cross-model review, and best-of-N runs.
imlazy – Tier-aware task router and skill-based workflow system for Claude Code that classifies tasks into Quick/Standard/Heavy/Vault/Learn tiers and routes them through the appropriate execution workflow with adaptive disambiguation.
agmem – Local-first project memory CLI for coding agents (Claude Code, Codex, Cursor, Aider). Indexes repo structure into greppable JSONL, retrieves task-scoped context via BM25 with optional local embeddings, and tracks staleness via source hashes. No LLM calls, no server. Apache-2.0.
Codex How To – Engineering-first OpenAI Codex guide with nine installable skills, measurable development and review loops, orchestration patterns, and a review-first Living Wiki.
Boost – Free CLI that reduces terminal and CI output before it reaches Cursor, Claude Code, and Codex, typically saving 60–90% of log tokens with reversible retrieval and local performance reports.
harness-ai-kit – Package manager for AI agent assets (skills/CLIs/MCPs/loops). Install, lock, and sync skills across Codex, Claude Code, Cursor, Kiro with SAT-based dependency resolution and SHA-256 lockfile. 23 curated enterprise skills included. Available on PyPI.
Monitor Token – Open-source Windows/macOS desktop widget for local Codex token usage, daily/weekly/monthly breakdowns, and weekly quota reset tracking.
NestMux – Desktop workspace for Windows, macOS and Linux that runs Claude Code, Codex, Gemini, Copilot and OpenCode side by side in a resizable pane grid, each pane with its own account and isolated HOME. Includes git worktrees, an integrated diff viewer, MCP server management and shared team workspaces. Free tier available.
Mydentify AI Model Cost Calculator – Local-first, no-signup calculator for estimating monthly AI model API costs from token usage, cached input, requests, and fixed per-request charges.
LWC – Open-source proactive project-memory CLI for AI coding agents, with source-grounded durable knowledge, SQLite/FTS5 retrieval, atomic updates, optional document/code graphs, and MCP/Skill/Hook integration across supported tools.
MCP Lens – Open-source DeepSeek Harness plugin and GitHub Action for progressive MCP tool discovery; its published 1,000-tool benchmark reduces the model-facing MCP interface from 1,000 schemas to 2 and the component schema JSON from 647,962B to 1,114B.
DSH Studio – Open-source cross-platform desktop host for installing, running, health-checking, and supervising DeepSeek Harness locally.
Forgeo – Scheduled, agent-driven software factory that works a JSON backlog with your coding agent (Claude Code, Codex, opencode), committing each result to the default branch and running a refactoring pass when the backlog is empty.
Vibe-Coding Prompt Template – Five-step planning workflow (deep research, PRD, tech design, AGENTS.md, build) with an npx vibeworkflow CLI that interviews you and writes the docs your coding agent reads. Free and MIT licensed.
sol-simplify – One-markdown-file skill for Claude Code and Codex that stops agents from inventing bureaucracy (approval gates, governance docs, validators for the validators) around their own work. Measured on gpt-5.6-sol: ceremony scored 0 in 10 of 11 benchmark skill runs vs 4-6 for baselines, every score cited to a line in committed raw output.
PM Skills – Open-source (MIT) library of 1,153 Agent Skills as plain-markdown SKILL.md files, covering PRDs, incident postmortems, code review checklists, RFCs, and 35 professions. Installs into Claude Code, Cursor, Windsurf, Zed and 10 more tools with one command, ships an MCP server, and runs free in a browser playground. Install: npx pm-claude-skills add.
Agent Operator Score – Local-first CLI (aos) that scores how well you operate Claude Code, Codex, and Grok CLI from your own session transcripts, plus a controlled-run assessment suite with a hidden verifier. No model calls in review mode; nothing leaves your disk.
Agent Coordinator – Per-user Codex skill that represents dependency-aware tasks as bounded work graphs, records revisioned local state, reconciles uncertain work before retry, reruns planned checks at closeout, and can use optional specialists or run inline. Free and open source (MIT).
Meanwhile – Claude Code/Cursor/Copilot/VS Code status line that shows a genuinely useful tip, or an occasional disclosed sponsor line, while the agent works. Sponsor earnings are split 50/50 with the developer, paid via PayPal. Never reads code or prompts. Free and open source.
AllPet – Open-source (MIT) desktop pet that watches your AI coding agents (Codex, Claude Code / Desktop, DeepSeek Harness, Grok) from local session logs — no accounts or API keys — and wakes the original session on click. Cross-platform status/watch CLI.
ManualMode – Skill and MCP workflow for hands-on coding practice alongside AI agents: suggest small repository tasks, reserve one for the developer, and track practice progress. Includes a free tier.
trailhead – Runs a large project as a map of decision tickets on GitHub Issues, resolved one at a time via a discuss→plan→execute→verify engine with atomic commits, an optional cross-AI plan review, and throwaway UI prototypes. Works with Claude Code and Codex; needs only an authenticated gh CLI. Open source, free.
Wave – Lets AI coding agents owned by different people talk to each other. Create a channel, paste one prompt into each agent (Claude Code, Codex CLI, Cursor, or anything with a shell), and the agents exchange messages while their humans watch and steer from the browser. No install, no account. Open source (MIT) and self-hostable.
myc – Open-source local memory and task queue for AI coding agents (Claude Code, Codex, opencode, Kimi): decisions survive context compaction, memory is anchored to code and follows it, everything stays in one SQLite file with no API key.
sidegrade – Reads your local coding-agent usage (Claude Code, Codex, Hermes, OpenCode) and shows which model gives you the same intelligence for less — scoring every model on the Artificial Analysis Intelligence Index and re-pricing your own token mix. 100% local, no account, MIT-licensed. Run with npx sidegrade; on npm.
MySpec – Interactive AI architect that interviews developers and generates structured 4-file spec bundles with MCP server integration.
npx @runapi.ai/mcp. Free catalog tools for model discovery and pricing; authenticated tools for media generation and LLM chat.https://markovo.net/mcp) plus local stdio server sandboxed to a dedicated root; free tier included.npm install -g prompt-to-asset.55 followers · starred Oct 2026
46 followers · starred Feb 2026
124 followers · starred Sep 2026
145 followers · starred Sep 2025