Agent Skills

Recursive Mode

recursive-mode

Orchestrates recursive repository workflows: phased runs, locks, audits, and durable memory updates for long-running agentic tasks.

Create Item (Task Orchestrator)

task-orchestrator

Turn conversation context into MCP work items: infer type/priority, anchor into containers (Bugs, Features, Tech Debt), create single items or work trees, and p

Agent Harness — Multi-session Project Framework

agent-harness

Provides a command-driven framework for planning sprints, running coding sessions, and orchestrating parallel subagent work for long-running projects.

Maestro: Architectural Governance

claude-code-maestro

Governance protocol that turns an agent into an architecture-focused conductor, enforcing init checks, planning, and verification for complex code tasks.

Run Planner (Run-Plan)

specs.md

Discover pending work items in the workspace and propose run grouping options (single, batch, wide) with dependency-aware ordering and user-preference learning.

Group Collaboration — Four-Tier Architecture

llm-autonomous-agent-plugin-for-claude

Guidelines and communication patterns for coordinating four agent groups (analysis, decision, execution, validation) to enable reliable feedback loops, knowledg

Compose Agents

orxhestra

Orchestrate multi-agent workflows with SequentialAgent, ParallelAgent, and LoopAgent for pipelines, parallel analysis, and iterative review loops.

Project Management (PM) Playbook

genie

Structured project-management skill that triages backlogs, prioritizes work, assigns owners, dispatches specialists, and produces status reports across Copilot,

Dispatching Parallel Agents

orchestrator-supaconductor

Pattern for splitting independent investigations into parallel agents — dispatch one focused agent per independent problem domain to speed up debugging and repa

ds-fix — data-science mid-analysis fixer

edwinhu

Orchestrates diagnosis and targeted fixes mid-analysis: diagnose root cause, apply fixes with output-first verification, and update project learnings.

cmux Control (claw-mux)

claude-code-zero

Control cmux terminal topology and I/O: split panes, send commands, read output, monitor services, and show sidebar metadata.

ScrollClaw

scrollclaw

Orchestrates end-to-end AI-generated UGC video production: persona research, first-frame creation, A-roll/B-roll generation, assembly and virality scoring.

Fabric Ops

fabric-skills-settings

Operational skill for managing Microsoft Fabric data platforms: run VACUUM, orchestrate DAGs, update inventory, and perform maintenance routines and GDPR-safe p

Fix-Review Loop

genie

Dispatch a dedicated fixer subagent to address FIX-FIRST gaps, re-run a separate review up to 2 loops, and escalate to a human if unresolved.

SwarmClaw CLI — Agent Fleet Manager

SwarmClaw

CLI skill to manage SwarmClaw agent fleets: agents, tasks, chats, schedules, memory, goals, connectors, and system health via a JSON-friendly CLI.

Codeagent Wrapper

stellarlinkco

Run and orchestrate codeagent-wrapper tasks across multiple AI backends (Codex, Claude, Gemini, OpenCode) for parallel code analysis, large refactors, and multi

Polymath

buiphucminhtam

A production-grade thinking partner that researches, analyzes, and helps decide what to build before execution — ideal for pre-flight consultations and complex

gh-issues — GitHub Issue Orchestrator

opensoul

Orchestrates fetching GitHub issues, confirming targets, and spawning code-fix sub-agents to create PRs and handle review comments via the GitHub REST API.

Turkey Build

turkey-build

Multi-agent app builder and orchestrator: bootstraps projects, creates feature branches, and runs runtime-verified builds with a PM conductor and QA gates.

Task Orchestration (Copilot Loop)

awesome-omni-skill

Drive repository work one task at a time using a single authoritative .copilot-todo.yaml and a plan→execute→iterate loop with tools to init, next, update, and c

RePrompter

reprompter

Turn messy prompts into structured, scoreable prompts or multi-agent plans; also reverse-engineer great outputs into reusable prompts.

Upstash Workflow SDK (JS)

workflow-js

Guidance and examples for using the Upstash Workflow SDK to define, trigger, and manage serverless workflows with QStash-backed reliability.

Matrix Code Review

claude-matrix

A 5-agent, context-aware code review pipeline for comprehensive PR and staged-change reviews, including security detection, blast-radius analysis, triage, remed

Load Context

rosetta

Ensure an agent loads the most current execution context and prep steps (bootstrap, docs, tools) before planning or acting.

TICK.md

tick-md

Git-backed, Markdown-native task coordination CLI and MCP server for multi-agent workflows — claim tasks, avoid collisions, and keep an audit trail.

Team Builder

claude-skill-registry

Designs high-performing team structures using organizational psychology and can create new skills on-the-fly to fill capability gaps during team planning.

SDD Delta Spec Writer

gentle-ai

Compose structured SDD delta specs: generate ADDED/MODIFIED/REMOVED requirements and testable Given/When/Then scenarios for change proposals (executor-only).

Test Orchestrator Skill

comeonoliver

QA orchestration skill: defines test strategy, assigns unit/integration/E2E suites, enforces coverage thresholds, and coordinates CI test execution.

YCE (Code Enhance & Search)

yce-enhance

Agent skill that enhances fuzzy code-related requests and/or locates implementation in a codebase; outputs structured XML results for downstream processing.

Building Agent Systems

code-abyss

Reference guide for designing single- and multi-agent LLM systems, RAG pipelines, security hardening, and prompt engineering for production agents.