All TypeScript packages
Every MIT-licensed TypeScript package across the reaatech AI catalog. Filter by category or repo, search by name, summary, or slug, and jump straight to the package detail or its npm listing. Packages that haven’t shipped to npm yet still appear here with a pending npm badge so the catalog reflects the actual surface — see solutions for assembled recipes.
Category
Repo
310 packages · page 1 of 13
@reaatech/a2a-reference-auth
Pluggable authentication and authorization strategies for A2A agents, providing `NoneStrategy`, `ApiKeyStrategy`, `JwtStrategy`, and `OAuth2Strategy` classes plus a `extractScopes` utility function. Designed for use with the A2A agent framework.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-cli
Description pending.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-client
A type-safe A2A client SDK that provides a class (`A2AClient`) for task submission, agent discovery, and SSE streaming against A2A-compatible servers. It handles agent card caching, retries with exponential backoff, and supports both synchronous task management and async event streaming via `sendSubscribe`.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-core
Zod schemas, TypeScript types, error classes, and a signature verification function for the A2A (Agent-to-Agent) protocol, providing canonical validation and parsing for task lifecycle messages, agent cards, and security schemes.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-mcp-bridge
A bidirectional protocol adapter that exposes MCP tools as A2A skills via `McpToolAdapter` (producing an enriched Agent Card), and wraps an A2A agent as an MCP server via `A2aAsMcpServer` (starting a stdio or SSE transport).
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-observability
Pino-based structured logging with correlation IDs and OpenTelemetry abstractions (tracers, meters, spans) for A2A agents, providing `createLogger`, `withCorrelationId`, `getTracer`, `getMeter`, and `withTaskSpan` functions with no-op defaults that can be replaced by a custom `TelemetryProvider`.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-persistence
Provides in-memory, filesystem, Redis, and Postgres implementations of a `TaskStore` interface for persisting A2A (Agent-to-Agent) tasks, each exposed as a class constructor.
- status
- published
- published
- 12 days ago
@reaatech/a2a-reference-server
A server framework that implements the A2A (Agent-to-Agent) protocol, providing Express 5 and Hono adapters that expose JSON-RPC 2.0 endpoints, SSE streaming, health checks, rate limiting, and push notifications for agent communication.
- status
- published
- published
- 12 days ago
@reaatech/agent-auth-proxy-client
Two typed HTTP clients for the agent-auth-proxy server: `AgentClient` exchanges an API key for a JWT and makes proxied requests to third-party APIs on behalf of a user, while `AdminClient` manages users, agents, grants, and tokens. Both decode server errors into typed error classes (`AuthError`, `ScopeError`, `ValidationError`, `UpstreamError`, `AppError`).
@reaatech/agent-auth-proxy-core
Shared Zod schemas, error classes, and TypeScript types for OAuth2 proxy request validation, scope management, and error handling, exported as framework-agnostic primitives that only depend on `zod`.
@reaatech/agent-auth-proxy-server
Fastify plugin and CLI that implements an identity-aware proxy server for agent-to-service communication, handling API key auth, OAuth2 with PKCE, JWT issuance, scope enforcement, credential injection, and audit logging. Exports `buildApp()` and `start()` functions, requires a PostgreSQL database and pairs with `@reaatech/agent-auth-proxy-core` for shared schemas.
@reaatech/agent-budget-cli
A CLI that installs a global `agent-budget` binary for managing agent budgets — defining budgets, checking remaining spend, listing active scopes, generating spend reports, simulating pre-flight checks, and manually resetting budgets.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-engine
A budget enforcement engine that provides pre-flight cost checks, real-time spend recording, and per-scope state machine transitions (Active → Warned → Degraded → Stopped) with auto-downgrade to cheaper models and expensive tool filtering. Exports a `BudgetController` class that orchestrates spend tracking, pricing, and policy evaluation, and pairs with `@reaatech/agent-budget-spend-tracker` and `@reaatech/agent-budget-pricing` at runtime.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-llm-router-plugin
A Fastify plugin that adds budget-aware routing to LLM Router, filtering model candidates by remaining budget and blocking requests when budgets are exhausted. It installs as the highest-priority strategy in the routing chain, evaluating before semantic and latency-based routing, and requires `@reaatech/agent-budget-engine` and `@reaatech/agent-budget-spend-tracker` at runtime.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-middleware
Express/Fastify middleware and a direct SDK (`BudgetInterceptor`) that enforces per-scope budgets on agent requests by checking estimated cost, auto-downgrading models, filtering expensive tools, and recording spend — all before the LLM call reaches your application code.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-otel-bridge
Converts OpenTelemetry GenAI spans into budget-tracked spend entries in real time by feeding span attributes to a `BudgetController`. Provides a `SpanListener` class that you call with span data from any OTel-compatible tracer.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-pricing
A pricing engine that computes LLM token costs using built-in or custom pricing tables, with LRU-cached lookups and model name normalization. Exports a `PricingEngine` class with `computeCost`, `estimateCost`, and `lookup` methods, plus a `ModelNormalizer` for alias resolution.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-spend-tracker
A circular-buffer-based in-memory spend tracker that records cost events and provides O(1) per-scope spend lookups, sliding-window rate calculations, cost projections, and anomaly detection—exposed as a `SpendStore` class with zero external dependencies.
- status
- published
- published
- 5 days ago
@reaatech/agent-budget-types
Zod-validated TypeScript types, enums, and error classes defining budget scopes, policies, enforcement actions, spend entries, and state transitions for the agent-budget-controller ecosystem. Exports interfaces, enums, and Zod schemas with zero runtime dependencies beyond `zod`.
- status
- published
- published
- 5 days ago
@reaatech/agent-chaos-adapters
Framework adapters for injecting fault-tolerance testing into LangChain, LlamaIndex, Vercel AI SDK, or any custom agent with a tool-call interface, using a shared interceptor pattern that wraps tools without modifying agent code. Each adapter exports a framework-specific wrapper (e.g., a tool wrapper for LangChain, a tool spec wrapper for LlamaIndex) that pairs with the `@reaatech/agent-chaos-core` engine at runtime.
- status
- awaiting publish
@reaatech/agent-chaos-cli
A CLI that validates, generates, and runs chaos engineering scenarios for AI agents, providing `init`, `generate`, `validate`, and `run` commands for managing fault-injection test scenarios defined in YAML or JSON.
chaoschaos-engineeringfault-injectioncliagent
- status
- published
- published
- 1 month ago
@reaatech/agent-chaos-core
Middleware-based fault injection engine for agent systems that intercepts tool calls and injects latency, timeout, malformed output, token exhaustion, and other failure modes via a single `engine.intercept()` call, with scenario-driven YAML/JSON configuration, probability-based fault selection, and seeded randomness for reproducible testing.
chaoschaos-engineeringfault-injectionagentllmresilience
- status
- published
- published
- 1 month ago
@reaatech/agent-chaos-e2e
End-to-end test suite that validates the full agent-chaos pipeline—scenario loading, schema validation, fault injection, engine event recording, CLI execution, and cross-package integration—against mock agent implementations.
- status
- awaiting publish
@reaatech/agent-chaos-observability
Structured logging, metrics collection, OpenTelemetry tracing, and report generation for chaos engineering experiments, exposed as a set of classes (`MetricsCollector`, `Tracer`, `ReportGenerator`) that consume events from `@reaatech/agent-chaos-core`.
- status
- awaiting publish