@fileoverview Core conversation loop engine for open-multi-agent. {@link AgentRunner} is the heart of the framework. It handles: - Sending messages to the LLM adapter - Extracting

mnemon.moth ----------- Auto-instrumentation engine. On mnemon.init(), the moth wakes up, scans the environment for installed agent frameworks, and patches each one at the right ex

Mnemon -- execution caching and system learning for AI agents. Zero code changes. One import line. Gets smarter with every run. Saves tokens on every cache hit. Usage: import mnemo

Zod definitions for workspace structure, UI state, and a function to create a fresh workspace instance. Reuse in applications managing AI canvas workspaces.

Builds typed execution receipts with token usage, validating schema; guards used for receipt creation.

Builds a case-insensitive regex pattern matching @mentions of a name for triggering agent invocation.

Executes evaluation test cases against targets with configurable concurrency, scoring, progress tracking, and cost/token metrics collection.

Default session key used for chat conversations. This value is used when no custom session key has been configured. It follows the format: agent:<agentId>:<sessionName> /

Renders 3D scene component "game" = "frontend" "room" "objects"di "space graph", "displayed via 1. Perhaps a "3D" main score 0.

AI browser/agents. Manages web page state. Persisted at end.

claude_node.exceptions — Typed exception hierarchy for Claude CLI controller. All exceptions inherit from ClaudeError (which inherits from RuntimeError), allowing callers to catch

Factory function that builds and configures a Flask app for AgentScope, including secret key generation.

Provides async browser automation using Playwright, supporting page interactions, authentication, and AI agent workflows.

@fileoverview Shared OpenAI wire-format conversion helpers. Both the OpenAI and Copilot adapters use the OpenAI Chat Completions API format. This module contains the common convers

@fileoverview {@link AISdkAdapter} — bridge to the Vercel AI SDK (`ai` + `@ai-sdk/*`). When {@link AgentConfig.adapter} is set, {@link Agent} skips {@link createAdapter} and uses t

Defines provider configs, model profiles, rate limits, tool-calling options for LLM routing.

Manages episodic memory items, providing creation, search (BM25), and embedding updates. Ideal for storing and retrieving time‑ordered event snippets tied to an agent.

Agent: unified agent class

@fileoverview {@link AgentBackend} that drives an external coding agent over the Agent Client Protocol (ACP). OMA takes the ACP **client** role: it spawns the agent (a local CLI su

Generates plain‑text documentation of all Chakra UI components for LLM consumption, cached for one hour.

Constructs a main coding prompt using dedented template with shadcn documentation context for frontend development guidance.

Core interface definitions and reference implementations for agent runtime, messaging, models, tools, workbenches, memory, and tracing. The foundational layer all other packages bu

Manages] F: validates

Defines the Bento class for packaging ML models with metadata and artifacts.

Workbench that wraps MCP (Model Context Protocol) servers to dynamically list and call their tools. Supports stdio and SSE server transports.

Defines protocol constants and command enums for an AI assistant SDK, enabling communication between AI models and Android device capabilities like camera, SMS, location, and notif

Shared helper for reading Claude Agent SDK message streams. Both the judge and Monte Carlo layers consume the same `query()` message stream (assistant text blocks + a terminal resu

Abstract base class defining the interface and core attributes (role, goal, backstory, tools, cache, execution parameters) for all agents in the CrewAI framework.

Normalizes AI model configuration data for provider-specific usage.

Server‑side helpers to fetch, list, cancel, and update AI generations with workspace access enforcement.

Schemas for various memory structures, context window overview, and archival/recall memory handling in Mirix.

Module exposing: getLanguageModel, listProviders, gatewayToken, hasGateway, hasAiBinding, gatewayProviderName, providerKey

Provides a Next.js request handler that wraps the Giselle AI engine, parses JSON/form bodies, routes to configured endpoints, and optionally runs after‑function logic.

eslint-disable @typescript-eslint/no-explicit-any */ * TypeScript wrapper for the GeminiNano native Expo module. At build time the native module is linked automatically by Expo. We

Mnemon moth — LangChain integration. System 2 EME via per-step patching: RunnableSequence.invoke is patched to iterate each step individually. Each step is a segment. If a step's o

Generates HTTP headers required for AI Gateway requests, including Stripe customer identifiers for usage attribution and billing.

compressors.py — Four compression strategies for the Engram Benchmark. Strategies: 1. EngramCompressor — LLM Observer + Reflector (Layer 6 Engram) 2. RuleCompressor — claw-compacto

High-level single and multi-agent APIs including AssistantAgent, UserProxyAgent, AgentTool, and group chat primitives. Built on autogen-core.

Decorator-based registry to attach validation functions for tool calls and runtime validator that checks arguments before execution.

Renders a terminal dashboard for a captured OpenMultiAgent run.

Single source of truth for default embedding and reranker model identifiers and dimensions used by schema and registry helpers.

Collects README, manifest and hotspot data, builds prompts for summarizer and weighter agents, and integrates their outputs into repository intelligence.

Builds LLM judge scorer: constructs prompts, parses verdicts, manages aborts/deadlines for eval.

Registers Gemini models with their capabilities, pricing, and defaults for integration into the LLM system.

SQLAlchemy model for tools that LLMs can invoke, with organization scoping and uniqueness constraints.

Mnemon Execution Memory Engine (EME) — v2 Generalised plan cache for any expensive recurring computation. System 1: exact fingerprint match — zero LLM, sub-millisecond System 2: pa

Agent identity, rules, memory, tools, skills, and hooks are all version-controlled files, so agents can be forked, branched, and diffed. Strong convergence with the OMC agent catal

Defines Mistral models such as Codestral with capabilities, pricing, and defaults for the LLM framework.

Combines multiple retrieval results using Reciprocal Rank Fusion to improve RAG answer quality via a query pipeline.

Converts user‑provided model identifiers into provider‑specific formats for various LLM services, handling aggregators, Anthropic, DeepSeek, and custom cases.