agent▌

Persistent, searchable memory system for agents to store and retrieve architectural decisions, patterns, and project knowledge. \n \n Provides four core MCP tools: memory_search for querying by text/type/tags, memory_write to record decisions and patterns, memory_read to retrieve specific entries, and memory_stats for usage analytics \n Runs as an MCP server that syncs with project documentation, enabling long-term knowledge retention across agent sessions \n Includes a standalone dashboard (por

Design and build AI agents with tools, memory, and multi-step reasoning capabilities. Covers ChatGPT, Claude, Gemini integration patterns based on n8n's 5,000+ AI workflow templates.

Delegate coding tasks to Codex, Claude Code, Pi, or OpenCode agents via background bash processes. \n \n Supports four coding agents (Codex, Claude Code, Pi, OpenCode) with agent-specific execution modes: PTY required for Codex/Pi/OpenCode; Claude Code uses --print --permission-mode bypassPermissions without PTY \n Background execution with session monitoring via process tool actions (log, poll, write, submit, kill) for long-running tasks \n Designed for feature building, PR reviews, refactoring

Slack automation CLI for reading, searching, sending, and managing messages, channels, and files. \n \n Supports message operations: fetch single messages or threads, browse channel history, send/edit/delete messages, add/remove reactions, and draft rich-text messages in a browser editor \n Download attachments (snippets, images, files) automatically; search messages and files across channels with date and user filters \n Create channels, invite users (including external), open DMs or group DMs,

Coordinate specialized agents to solve complex problems through orchestrated workflows and collaborative reasoning. \n \n Supports four core orchestration patterns: sequential task chains, parallel execution, hierarchical delegation, and consensus-based debate \n Includes templates and examples for CrewAI, AutoGen, LangGraph, and OpenAI Swarm frameworks \n Provides utilities for agent communication (message brokers, shared memory), workflow management (execution, optimization, monitoring), and p

Multi-layer memory architecture for agents: short-term context, long-term vector storage, and retrieval optimization. \n \n Covers seven memory types: short-term (context window), long-term (vector stores), working, episodic, semantic, and procedural memory, each suited to different information patterns \n Provides three core patterns: memory type selection, vector store choice, and chunking strategy to maximize retrieval accuracy \n Highlights critical retrieval challenges: contextual chunking,

$23

Operational controls for long-lived agent workloads with observability, security, and lifecycle management. \n \n Covers four operational domains: runtime lifecycle (start, pause, stop, restart), observability (logs, metrics, traces), safety controls (scopes, permissions, kill switches), and change management (rollout, rollback, audit) \n Enforces baseline controls including immutable deployment artifacts, least-privilege credentials, environment-level secret injection, and hard timeouts with re

Build, run, and debug iOS apps on a booted simulator with UI interaction and log capture. \n \n Discovers booted simulators, sets session defaults (project path, scheme, simulator ID), and builds/runs the current iOS project via XcodeBuildMCP tools \n Provides UI inspection ( describe_ui , screenshot ), interaction commands ( tap , type_text , gesture ), and element targeting by ID, label, or coordinates \n Captures simulator logs and console output with start/stop controls; includes build failu

You are a background coding agent running inside a Replicas cloud workspace (a remote VM). This guide covers capabilities and best practices specific to this environment.