AgentWorkforce/relay

Agent Relay coordinates registered agents in real time using channels, DMs, threads, reactions, inbox checks, and worker spawn or release. Use it when your session is already in a Relay workspace and needs to communicate or coordinate with peers.

Agent Relay coordinates multiple Codex agents through shared messaging, handoffs, and status tracking. Use it when workers need peer-to-peer updates, ACK/BLOCKED/DONE signals, or a shared workspace.

Animation Review checks animation and motion code against a strict craft bar focused on timing, easing, origin, performance, and accessibility. Use it when reviewing UI motion changes and deciding whether they should be approved.

Browser Testing With Screenshots automates Chrome interactions and screenshot capture to verify web app UI flows. Use it when you need visual confirmation of layouts, forms, or localhost behavior during development.

Choosing Swarm Patterns helps you pick the right Agent Relay orchestration pattern for coordinating multiple AI agents. It covers the core and specialized swarm patterns, with decision guidance and YAML or fluent builder examples.

Choosing Swarm Patterns helps you pick the right Agent Relay orchestration pattern for coordinating multiple AI agents. Use it to choose and configure YAML or builder-based workflows for fan-out, pipeline, hub-spoke, consensus, mesh, handoff, cascade, dag, debate, and hierarchical patterns.

Command Palettes covers the React patterns for Cmd+K menus, including arrow-key selection, keeping the active item in view, shortcut matching, and avoiding re-render loops from unstable references.

Creating Agent Skills explains the SKILL.md format for portable agent capabilities across Codex CLI, GitHub Copilot, and Amp. Use it when authoring skills with the required frontmatter, directory structure, and validation rules.

Creating Claude Agents helps you build or improve Claude Code agents with the right file structure, frontmatter, persona, tools, model selection, and schema validation. Use it when creating agents or fixing agent format issues.

Creating Claude Hooks explains how to build Claude Code hooks with the right executable format, event handling, JSON I/O, exit codes, and package structure. Use it when creating, troubleshooting, or publishing a hook package.

Creating Claude Rules helps you create and fix .claude/rules/ files with the correct paths frontmatter, quoted glob patterns, and Claude-specific formatting. Use it when migrating Cursor rules or organizing project-scoped instructions.

Creating Skills helps you write new Claude Code skills or improve existing ones with clear structure, discoverability, and examples. Use it when a technique needs to be reusable, scannable, and easy for Claude to find and apply.

Debugging WebSocket Issues helps diagnose invalid frame header, RSV1, and unexpected close errors in WebSocket servers. Use it to trace upgrade conflicts, compression mismatches, and raw HTTP written to sockets.

Deploying To Staging Environment coordinates synchronized staging pushes across relay, relay-dashboard, and relay-cloud. Use it when you need to verify cross-repo changes or trigger the staging GitHub Actions workflow.

Design Engineering encodes Emil Kowalski's philosophy for UI polish, component design, and animation decisions. Use it when reviewing or designing interfaces that should feel precise and thoughtful.

Frontend Design creates distinctive, production-grade interfaces for components, pages, and applications. Use it when you need polished frontend code with a clear visual direction instead of generic AI-style layouts.

GitHub OAuth Nango Integration implements separate Nango connections for GitHub login and GitHub App repo access. Use it when linking user identity to installations, handling webhooks, and syncing repository data.

PRPM JSON Best Practices explains how to structure prpm.json manifests for publishing packages, including required fields, tags, multi-package repos, collections, and activation settings. Use it when creating or maintaining PRPM package metadata.

Relay 80 To 100 Workflow defines agent-relay processes that validate features end to end before merge. It is used for workflows that need repair loops, test-fix-rerun gates, and fresh-eyes review before commit.

Relay 80-100 Workflow defines agent-relay patterns that validate features end to end before merging. It uses repair-before-failure gates, test-fix-rerun loops, and sequential Claude-then-Codex review passes to ensure green acceptance.

Relay Workflows helps build multi-agent workflows with @relayflows/core using DAG steps, agent coordination, and output chaining. Use it for review/fix loops, verification gates, and chat-native or pipeline-style orchestration.

Relayfile Setup configures relayfile so agents can read and write provider files through a local mount. Use it to create a workspace, verify sync and writeback, and hand the mount off to a local or remote agent.

Relayfile Setup mounts provider data as files so agents can read and write through a local directory. Use it to connect a workspace, verify sync, and troubleshoot writeback or cloud-mount issues.

Review Fix Signoff Loop coordinates repeated review, fix, and validation cycles with fresh agent context until independent signoff agents both approve the result. It is used for high-stakes workflows that need deterministic gates, repair passes, and final signoff reporting.

Review Fix Signoff Loop defines a workflow for implementation tasks that must cycle through review, repair, validation, and fresh-context signoff until independent reviewers agree the work is complete. Use it for high-stakes agent workflows with repairable gates, dual-verdict contracts, and PR signoff reporting.