Skip to content
/
§ a8

Cross-runtime harness

One source compiled to outputs for Claude Code, Codex, Gemini CLI, Cursor, etc.

69 primary frameworks · 50 lower-confidence entries

claude-mem (thedotmack)
claude-mem
★ 78k
Tier A

Background worker service captures every tool call as an observation, AI-compresses sessions, and auto-injects relevant past context using a 3-layer…

pi (badlogic/earendil)
pi-coding-agent
★ 55k
Tier A

A minimal, hackable, multi-provider terminal coding agent that adapts to your workflows via npm-installable TypeScript Extensions and Markdown Skills — without…

Agent Skills (Addy Osmani)
agent-skills-addyosmani
★ 46k
Tier A

Encodes senior-engineer software development lifecycle as 23 auto-routed skills and 7 slash commands for any AI coding agent.

wshobson/agents Plugin Marketplace
wshobson-agents
★ 36k
Tier A

Single Markdown source for 83 domain-specialized plugins that auto-generates idiomatic artifacts for five AI coding harnesses.

TabbyML/Tabby
tabby
★ 34k
Tier A

Self-hosted AI coding assistant server (alternative to GitHub Copilot) with admin dashboard, RAG-based completions, and multi-IDE support.

Compound Engineering
compound-engineering
★ 17k
Tier A

Make each unit of engineering work compound into easier future work via brainstorm→plan→execute→review→learn cycles.

Qodo (PR-Agent)
qodo
★ 11k
Tier A

Open-source AI PR reviewer with single-call tool architecture, PR compression for large diffs, self-reflection quality gate, and cross-platform git provider…

Superset
superset
★ 11k
Tier A

macOS desktop app that eliminates context-switching overhead when running 10–100+ parallel AI coding agents, each isolated in its own git worktree, monitored…

worktrunk
worktrunk
★ 5.2k
Tier A

Make git worktree lifecycle (create, switch, list, merge, cleanup) as simple as branch operations, designed for managing 5-10+ parallel AI agents.

mirrord
mirrord
★ 5.1k
Tier A

Routes local process syscalls (network, files, DNS, env) through a live Kubernetes cluster pod so AI agents test against real infrastructure without deploying.

mini-swe-agent
mini-swe-agent
★ 4.5k
Tier A

Prove that a ~130-line agent with only a bash tool can achieve >74% on SWE-bench, and serve as a clean research baseline + hackable daily tool.

Micro Agent (Builder.io)
micro-agent
★ 4.3k
Tier A

Generate code that passes tests, nothing else — the smallest possible TDD agent that avoids the Roomba-under-the-table problem of general coding agents.

Mistral Vibe
mistral-vibe
★ 4.3k
Tier A

Mistral AI's first-party open-source CLI coding agent with configurable safety profiles, subagent delegation, SKILL.md custom skills, and ACP programmatic…

Flue
flue
★ 3.7k
Tier A

TypeScript server framework for building deployable headless agent services, with first-class Cloudflare Workers + Durable Objects support

cc-sdd (gotalab)
cc-sdd
★ 3.4k
Tier A

Portable Kiro-style SDD harness for 8 AI agents: discovery-routed spec creation with boundary-annotated tasks and autonomous subagent-per-task TDD…

Amplifier (microsoft)
amplifier
★ 3.1k
Tier A

Ultra-thin Python kernel (~2,600 lines) with formal module protocol contracts and a Git-based bundle marketplace, providing a Linux-kernel-style extensible AI…

wshobson/commands Slash Command Collection
wshobson-commands
★ 2.5k
Tier A

57 slash commands (15 multi-agent workflows + 42 single-purpose tools) for Claude Code, now superseded by the wshobson/agents plugin marketplace.

APM (Agentic Project Management)
apm-agentic-project-mgmt
★ 2.3k
Tier A

Manages complex multi-session software projects by coordinating specialized AI agents (Planner/Manager/Workers) with human-mediated message relay and…

TmuxAI
tmuxai
★ 1.8k
Tier A

Be the AI pair programmer that watches your tmux screen and helps with any terminal workflow — no shell wrappers, no workflow changes, just observe and assist.

Haft
haft
★ 1.3k
Tier A

Engineering reasoning governor that enforces First Principles Framing — frame, compare under parity, decide with falsifiable contracts, detect stale evidence —…

quint-code (Haft)
quint-code
★ 1.3k
Tier A

Governs AI agent execution through formal decision contracts, parity-enforced comparisons, evidence-decay scoring, and bounded WorkCommissions — ensuring…

Spec Kitty
spec-kitty
★ 1.3k
Tier A

Delivers a complete spec-driven agentic workflow across 17 AI agents with worktree isolation, lane-tracked parallel execution, and retrospective-based workflow…

deepagentsjs
deepagentsjs
★ 1.3k
Tier A

TypeScript port of LangChain's Python Deep Agents harness, with advanced generic type safety for agent configuration

agentflow (berabuddies)
agentflow-berabuddies
★ 1.3k
Tier A

Python DSL for building multi-agent DAG pipelines with >> operator, Jinja2 prompt chaining, fanout/merge parallelism, iterative LLM-as-judge cycles, and native…

Bhartendu-Kumar/rules_template
rules-template-stackable
★ 1.1k
Tier A

Cross-platform rule template for Cursor + Cline + Roo Code using symbolic links for a shared single source of truth with Agile-inspired SDLC workflow.

thClaws
thclaws
★ 1.0k
Tier A

One native Rust binary that is a complete sovereign AI agent workspace — GUI, CLI, webapp, scheduler, knowledge base, and multi-agent orchestration — without…

oh-my-agent
oh-my-agent
★ 1.0k
Tier A

Portable multi-agent harness that models an engineering team as specialized agents and projects them across 27+ AI tools from a single .agents/ source of truth

GET SHIT DONE (GSD)
get-shit-done
★ 943
Tier A

Prevent AI context rot by keeping the orchestrator lean (~15% context) and running all implementation in parallel fresh 200k-token subagent contexts.

AgentSys (avifenesh/awesome-slash)
awesome-slash-commands
★ 823
Tier A

Orchestrate everything around code-writing — task selection, branch management, review, PR, merge, memory — via structured pipelines with gated phases and…

context-space
context-space
★ 810
Tier A

Provides production-grade OAuth-secured MCP tool aggregation across 14+ external services, eliminating the hardest part of connecting AI agents to real-world…

kiro (jasonkneen)
kiro-jasonkneen
★ 665
Tier A

Verbatim Kiro IDE system prompt + 8 spec-driven development skills packaged as a Claude Code plugin for teams that want Kiro's methodology without the paid IDE.

flow-next
flow-next
★ 615
Tier A

Enforce spec-first task hygiene with fresh-context workers, source-tagged capture, and cross-model review gates to prevent context bleed and hallucinated…

GroundZero Package Manager (OpenPackage / opkg)
gpm-groundzero
★ 557
Tier A

Universal CLI package manager for AI coding agent configuration files (rules, commands, agents, skills, MCPs) with cross-platform conversion for 40+ tools.

MateClaw
mateclaw
★ 512
Tier A

Enterprise multi-agent server platform: always-on, multi-vendor failover, RBAC, approval gates, audit trail, and 8 IM channel adapters in a single Spring Boot…

SWE-ReX
swe-rex
★ 508
Tier A

Unified Python library for interacting with persistent bash sessions in any execution backend (local, Docker, Fargate, Modal).

Open Agent (Th0rgal)
open-agent-thorgal
★ 438
Tier A

Same as sandboxed-sh — self-hosted AI agent orchestrator (non-canonical slug).

sandboxed.sh
sandboxed-sh
★ 438
Tier A

Self-hosted orchestrator for AI coding agents with isolated workspaces, multi-runtime support, and Library-based configuration management.

SpecPulse
specpulse
★ 385
Tier A

Deploys an identical spec-driven workflow across 8 AI platforms using a CLI-first scaffold that the AI then populates.

Vet (Verify Everything)
vet-imbue
★ 385
Tier A

Catches intent-implementation mismatches in AI-generated code by cross-examining git diffs against the stated goal and conversation history.

crit
crit-review
★ 350
Tier A

Brings GitHub PR-style human inline review to AI agent output (plans, code, live apps, HTML) with persistent per-line comments and round-to-round diffs,…

Water
water
★ 288
Tier A

Python agent harness providing infrastructure (orchestration, resilience, observability, fallback chains) around any AI agent framework

OmniCoreAgent
omnicore-agent
★ 241
Tier A

Python production agent harness with parallel tool batching, structured observations, BM25 tool retrieval, signature loop detection, and modular production…

sd0x-dev-flow
sd0x-dev-flow
★ 157
Tier A

Reference implementation of harness engineering for Claude Code — hook-enforced dual review, sentinel-driven state machine, and fail-closed safety where the AI…

specs.md (AI-DLC)
ai-dlc-specs-md
★ 156
Tier A

Ships three selectable SDD methodologies (Simple/FIRE/AI-DLC) in one npm package so teams can graduate from lightweight to full lifecycle orchestration without…

Conductor (microsoft)
conductor-microsoft
★ 156
Tier A

Deterministic YAML-defined multi-agent workflow engine with Jinja2 routing, parallel execution, multi-model support, and a built-in real-time web dashboard.

hankweave
hankweave
★ 123
Tier A

Production operations runtime for executing frozen, long-horizon agentic programs (hanks) reliably, with single-threaded execution, git checkpointing, and…

HexAgent
hexagent
★ 122
Tier A

Gives any LLM a computer via a runtime-computer isolation protocol — the harness never shares its keys or config with the agent.

MetaSpec (ACNet-AI)
metaspec-acnet
★ 47
Tier A

Generates complete spec-driven toolkits for any domain from a single command, enabling the creation of domain-specific specification systems rather than…

shinpr/agentic-code
shinpr-agentic-code
★ 46
Tier A

AGENTS.md-based framework that enforces TDD, Plan Injection gates, and progressive skill loading across Cursor, Codex CLI, and Gemini CLI.

MemoryAgent
memoryagent
★ 38
Tier A

Implements memory as a plain text file managed by the agent's native Read/Write/Edit/Grep tools, with a structured analyze command that produces a 7-section…

Tessl
tessl
★ 38
Tier A

Steering tile that enforces spec-before-code methodology via versioned skills, always-on rules, and an evaluation harness with 9 graded scenarios.

Tessl SDD Tile
tessl-sdd-tile
★ 38
Tier A

Provides a skills+rules+evals tile for spec-driven development with one-question-at-a-time requirement gathering, explicit stakeholder approval gates, and…

Brood Box
brood-box
★ 36
Tier A

Wrap any coding agent in a hardware-isolated microVM with COW workspace snapshot, egress firewall, and interactive change review.

SWORDSwarm
sword-swarm
★ 24
Tier A

88-agent corporate-hierarchy orchestration system with Intel NPU hardware acceleration and multi-IDE support for enterprise-grade parallel AI task execution.

Pluqqy
pluqqy
★ 21
Tier A

Compose reusable context/prompt/rule components into named pipelines that activate with one command, enabling instant context switching between development…

AI Engineering Harness
adrielp-ai-engineering-harness
★ 16
Tier A

Translates a shared set of context-engineering patterns into native formats for 4 different AI coding tools via a Deno CLI installer.

Aigon
aigon
★ 10
Tier A

Orchestrates 7 different AI coding agent CLIs from a single Kanban workflow with per-feature worktree isolation, Fleet mode for parallel agent comparison, and…

OneBrain
onebrain
★ 10
Tier A

Give AI agents persistent memory, 34 skills, and personal calibration via an Obsidian vault so every session picks up exactly where the last one left off —…

Cline ACP
cline-acp
★ 9
Tier A

ACP protocol adapter that exposes Cline's coding capabilities to non-VS-Code editors like Zed.

SpecD
specd-sdd
★ 9
Tier A

Pre-assembles deterministic context packages for agents at each lifecycle step, combined with a multi-language code graph and customizable schema that defines…

aiignore-cli
aiignore-cli
★ 8
Tier A

One command to generate correct, security-researched ignore configurations for all AI coding tools in a project, with documented CVEs and bypass…

Ozzeron prompt-pack
ozzeron-prompt-pack
★ 6
Tier A

Prevent the recurring AI-generated technical debt patterns (duplicate artifacts, convention drift, scope creep) that accumulate regardless of which AI tool is…

vibe-stack (vibestackdev)
vibe-stack
★ 6
Tier A

Prevent AI coding assistants from generating insecure or broken Next.js 15 + Supabase code by injecting stack-specific constraint rules into every AI…

che-incubator demo-spec-driven-development-with-ai
ra-aid-che-incubator
★ 2
Tier A

Zero-friction template for spec-driven development in cloud workspaces: write a spec, run one command, get working code.

TokRepo
tokrepo
★ 0
Tier A

Open registry infrastructure for AI assets where agents can autonomously discover, install, and contribute reusable capabilities across sessions and platforms.

DocBrain
docbrain
Tier A
Factory Droid
droid-factory
Tier A

Full-platform AI software factory: hooks turn prompts into deterministic enforcement, Missions add structured orchestration, and Droid Computers provide…

Maestro
maestro-orchestrate
Tier A
Patchwork OS
patchwork-os
Tier A
Show 50 lower-confidence entriestier-b · tier-c · unknown · delta reports

These entries map to § a8 by tag but carry weaker evidence — fewer documented primitives, delta reports of absent skills, or marketing-only sites without a public repo. They're listed for completeness; treat them with appropriate caution.

Multica ★ 33k
multica

Open-source managed agents platform with 4 UI surfaces (web/desktop/mobile/CLI), squad routing, autopilots, and skill compounding — turns coding agents into org…

NanoClaw ★ 29k
nanoclaw

Run Claude agents securely in per-session Docker containers with multi-channel messaging (WhatsApp/Telegram/Discord) and credential vault isolation.

Planning with Files ★ 22k
planning-with-files

Enforces Manus-style persistent markdown planning on any AI coding agent via hooks that automatically re-inject plan state before every tool call.

Microsoft Agent Framework ★ 11k
ms-agent-framework

Production-grade Python/.NET SDK for building, orchestrating, and hosting multi-agent AI workflows on Azure Foundry.

CAI (Cybersecurity AI) ★ 8.8k
cai-cybersecurity

Lightweight framework for AI-powered offensive and defensive security automation, battle-tested in CTF competitions and real-world vulnerability discovery.

Claude Code PM (ccpm) ★ 8.1k
ccpm

Five-phase PRD-to-shipped-code skill using GitHub Issues as canonical task store with parallel agent execution via git worktrees.

Plano ★ 6.5k
plano

AI-native proxy built on Envoy that externalizes agent orchestration, LLM routing, observability, and safety as out-of-process middleware — agents are just HTTP…

Backlog.md ★ 5.6k
backlog-md

Structured task management for AI coding agents using per-task markdown files, MCP, and a web Kanban board — all inside the git repo.

Archestra ★ 3.7k
archestra

Enterprise AI platform providing centralized MCP registry, Kubernetes-native orchestration, dual-LLM security, and cost management for organizations adopting AI…

AI-DLC Workflows (AWS) ★ 2.4k
aws-aidlc-workflows

Cross-IDE adaptive software development lifecycle enforcement for AI coding agents, with mandatory audit trails and human approval gates at each phase.

centminmod/my-claude-code-setup ★ 2.4k
centminmod-cc-setup

Provides a memory-resilient Claude Code starter with dual git-shared + machine-local memory that survives CLAUDE.md resets, plus multi-provider delegation to Co…

AgentBay SDK ★ 1.1k
agentbay-sdk

Multi-language SDK for Alibaba Cloud's on-demand sandbox sessions across Browser, Desktop, Mobile, and Code execution surfaces.

Caliber ★ 1.1k
caliber

Generates and continuously maintains AI context files (CLAUDE.md, Cursor rules, AGENTS.md, Copilot instructions) with deterministic quality scoring and automati…

Mysti ★ 1.1k
mysti

VS Code extension that orchestrates 12 AI coding agent CLIs through a unified chat UI with 5-strategy Brainstorm Mode, 16 personas, and heuristic convergence de…

Sidecar ★ 1.0k
sidecar-marcus

Terminal TUI companion panel that provides real-time git diff viewing, unified conversation history from 10+ AI coding agents, task monitoring, and workspace ma…

spec_driven_develop ★ 866
zhu1090093659-sdd

6-phase cybernetics-inspired pipeline for large-scale AI transformations with S.U.P.E.R architectural health framework, adaptive drift control, and GitHub-nativ…

MyCoder.ai ★ 567
mycoder-ai

Autonomous parallel-executing coding agent that reads project context naturally without special setup files.

oh-my-opencode (opensoft) ★ 544
opensoft-oh-my-opencode

Multi-provider multi-agent harness for OpenCode with Sisyphus orchestrator routing work across Anthropic/OpenAI/Google/xAI agents based on task domain.

AgentOps (boshu2) ★ 369
agentops-boshu

SDLC control plane for coding agents: compounding corpus of decisions, learnings, and planning rules that makes each session smarter than the last.

BMAD-AT-CLAUDE ★ 235
bmad-at-claude

Ports the full BMAD agile agent workflow (10 named personas, planning+implementation phases) into Claude Code's native hook and subagent infrastructure.

Shep ★ 194
shep-cli

Runs multiple AI coding agents in parallel, each in its own git worktree, handling commits, PRs, and CI monitoring automatically.

terminal-bench-env (TermiGen) ★ 82
terminal-bench-env

3,500+ verified Docker terminal tasks + minimal ReAct BashAgent for evaluating and training terminal-capable AI agents.

OpenSpecUI (jixoai) ★ 72
openspec-ui-jixoai

Provide a visual web interface for OpenSpec workflows with PTY terminal, OPSX compose, project hooks, and static export.

skill-optimizer ★ 57
skill-optimizer

Runs AI agent skills in hermetically isolated Docker containers against deterministic graders to produce reproducible pass/fail results across a model × case ma…

FredAntB/Spec-Driven-Development ★ 57
spec-driven-fredantb

Conversational skill that interviews the user once, generates requirements/design/tasks, then creates identical Universal Instruction Blocks for every AI tool s…

mcp2skill ★ 51
mcp2skill

Converts any MCP server into a production-quality agent skill package following the agentskills.io specification, with real introspection, OAuth support, and up…

AgentTrace ★ 49
agenttrace-luoyu

Local-first TUI and CI-gate tool for retrospective analysis of AI coding-agent session cost, tokens, health, and latency across 15+ agent runtimes.

unslop ★ 44
unslop

Strips AI-isms (sycophancy, stock vocabulary, hedging stacks, em-dash overuse) from LLM responses while preserving technical accuracy via research-backed subtra…

openspec-schemas (intent-driven-dev) ★ 43
openspec-schemas-intent

Extends OpenSpec with domain-specific artifact workflows for BDD, event-driven systems, ADRs, and Linear project management.

LynxPrompt ★ 41
lynxprompt-geiserx

Self-hostable platform for creating, versioning, and distributing AI coding assistant configurations (CLAUDE.md, AGENTS.md, .cursorrules) across teams with ente…

Aurite Agent Verifier ★ 38
aurite-agent-verifier

Applies deterministic pattern-matched checks on AI-generated agent code to catch hallucinated tools, unbounded loops, and missing retry limits before code ships…

faf-cli ★ 29
faf-cli

Provides a git-native, IANA-registered YAML format for AI project context (project.faf) that bi-syncs with CLAUDE.md/AGENTS.md and scores 0-100% AI-readiness vi…

Agentify ★ 28
agentify

Compiles OpenAPI specifications into all agent interface formats (MCP server, CLAUDE.md, AGENTS.md, skills, .cursorrules, A2A card, CLI) so API owners don't man…

Indie Kit ★ 22
ind-kit

A Next.js 16 SaaS starter kit with embedded Claude Code skills that prevent AI agents from hallucinating incompatible patterns by encoding exact import paths, A…

Tracer ★ 20
tracer-issue-tracker

Dependency-aware JSONL issue tracker for AI agents with a ready-queue that surfaces only unblocked work.

openspec-spec-driven-superpowers ★ 14
openspec-spec-driven-superpowers

Bridges OpenSpec's lifecycle control with superpowers execution discipline via a review.md readiness gate and per-change execution mode selection.

House MCP Manager ★ 8
house-mcp-manager

Manages MCP server enable/disable state across Claude Code, Cursor, and Cline to prevent context window exhaustion from startup tool-loading.

codebase-pilot-cli ★ 6
codebase-pilot-cli

Combines codebase context packing, security scanning, workflow skills, and a persistent web dashboard into one zero-cloud tool that reduces per-prompt token cos…

PromptKit (ozzeron/prompt-pack) ★ 6
ozzeron-prompt-pack-promptkit

Prevents AI agents from creating duplicate artifacts by enforcing a mandatory reuse-before-create decision chain as a precondition for any code-writing action.

HarnessOS ★ 3
harness-os

SQLite-backed execution harness that gives AI agents task lifecycle management, lease-based concurrency control, workload-profile skill specialization, and Symp…

foomakers/pair ★ 3
ozzeron-foomakers-pair

Gives AI assistants the full team context — PRD, ADRs, guidelines, process — needed to execute a complete SDLC without context loss.

agentic-python-coder ★ 3
szeider-python-coder

A focused Python execution agent (LangGraph ReAct + persistent IPython kernel) that mandates constraint-independent verification before saving any solution, plu…

intellij-openspec ★ 2
intellij-openspec

Provide full OpenSpec lifecycle orchestration within IntelliJ IDEA with flexible AI routing to any of 28 detected tools or direct API providers.

prpack ★ 2
prpack

Packages PR diffs with full post-change file contents so LLMs can review code with context about unchanged lines, not just what changed.

AI Context Linter ★ 1
ai-context-linter

CI linter that validates AI coding context files (CLAUDE.md, .cursorrules, AGENTS.md) against 12 rules covering security, structure, and AI anti-patterns.

AI Context Templates ★ 1
ai-context-templates

Provides annotated fill-in-the-blank starter files for AI coding context (CLAUDE.md, .cursorrules, PRP) so developers don't write useless or harmful context fil…

oh-my-agent-skills ★ 1
lukasdias-oh-my-agent-skills

Rust TUI for browsing, searching, filtering, and copying skill names from .agents/ skill directories used by Claude/Opencode coding assistants.

openspec-schemas (kmhalvin) ★ 1
openspec-schemas-kmhalvin

Prevents context budget exhaustion and hallucination on complex brownfield changes via subagent delegation and mandatory session restarts.

Marmot ★ 0
marmot

Provides a single shell-pipeable CLI for AI generation, web search, and data enrichment across 25+ providers so agents can delegate external calls without burde…

reqtext ★ 0
reqtext

Git-native requirements management CLI generating dual human+AI documentation from a flat JSON hierarchy with built-in test traceability.