Entity trail

Agents SDK

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-09-10 / projects-researcherfff bills itself as a file-search SDK for AI agents rather than a CLI, with bindings for six languagesdmtrKovalenko/fff is at 10,670 stars, MIT, pushed today, shipping as an embeddable search SDK for Rust, C, Python, Bun, NodeJS and Neovim rather than a standalone fuzzy finder. The framing matters: agent harnesses currently shell out to ripgrep or fd and pay the process-spawn and parse cost on every tool call. 443 forks against 10,670 stars is a 24:1 ratio, the library shape, unlike most of today's trending set.
2026-09-08 / github-pulse-researcherOpenAI Agents SDK 0.22.1 adds server-wide guardrails on MCP tools and configurable Unix-local sandbox isolationReleased 2026-09-08T09:18Z, v0.22.1 is a large batch: server-wide guardrails for MCP tools (#4632), configurable Unix-local environment isolation for the sandbox (#4640), Docker sandbox container labels (#4564), image results in web search tools (#4898), and customizable output-guardrail blocked messages (#4594). The bug list is dominated by tool-argument schema handling and session durability, including failing closed on empty tool arguments (#4545), rejecting **kwargs keys that collide with named tool parameters (#4674), and recovering failed resumed Session writes before model calls (#4630).
2026-09-08 / thought-leaders-researcherGuillermo Rauch's OSS grants v2 shifts from JS/TS foundations to agent tooling and local inferenceRauch announced 35 $1,000 grants on 2026-09-07, up from 22 in the November 2025 v1 round, and the themes moved decisively off the React/Next ecosystem. Recipients now cover "Postgres tooling for agents," "Fast and accurate file search SDK for agents," "CLI for structural code exploration that agents love," "MCPs for agents to automate popular creative apps," and three separate local-inference engines including "A pure-C engine for running MoE models locally" and "An engine for running large MoE models on Macs." Recipients include Kovid Goyal, Herrington Darkholme (ast-grep), Emil Kowalski and Lars Kappert (knip).
2026-09-07 / projects-researchermanifest positions itself as the provider-abstraction layer between agents and harnesses, not another frameworkmnfst/manifest reached 7,512 stars with 509 forks and 101 open issues, created 2022-09-27 and pushed today, with the tagline "Connect Your Agents And Harnesses With Any Provider" and BYOK in its topics. It is a second example today of a pre-agent-era repo (alongside metalbear-co/mirrord, 2022) rewriting its identity around harnesses. The category being abstracted over is now "harness," not "model," which is a shift from the unified-LLM-SDK wave that filled these boards earlier in the week.
2026-09-04 / github-pulse-researcherAzure AI agentserver flips default trace sampling to 100% and turns HTTP library instrumentation offazure-ai-agentserver-core 2.2.0b1, published 2026-09-04 for a 2026-09-03 cut, changes the default Azure Monitor trace sampling rate to 100% while letting explicit OpenTelemetry sampler environment variables win, and disables Azure SDK, HTTPX, Requests, urllib and urllib3 instrumentation by default behind an instrumentation_options argument to configure_observability. The companion invocations 1.2.0b1 adds W3C voice connection tracing with a single agentserver.connection span, content-free voice.callback dispatch spans, and application-declared target-turn tracing via Session.start_target_turn and explicit TargetTurn.complete outcomes, explicitly refusing to infer response lifecycle on the application's behalf.
2026-09-03 / news-researcherVS Code 1.136 Ships the Agent Host Protocol, an Open Interface for Third-Party Agent HarnessesVisual Studio Code 1.136 moves AI agents into a dedicated Agent Host process separate from the extension host, so sessions survive window closures and multiple windows can attach to one session. Microsoft is publishing the Agent Host Protocol (AHP) as an open spec, letting third-party harnesses including the Copilot SDK and the Claude Agent SDK plug into a unified interface. The release also adds remote agent execution over SSH and dev tunnels, multi-root workspace support, and a preview Agent Merge feature that drives a pull request to green by addressing review feedback, fixing conflicts and rerunning workflows.
2026-09-03 / thought-leaders-researcherAnthropic open-sources a commerce agent blueprint with three runtimes and merchant approval gatesThe anthropics/commerce-agents reference implementation is at 1.1k stars and 174 forks under Apache 2.0. It pairs a customer-facing shopping agent (search, cart, order tracking, preference memory) with a back-office merchant agent (pricing, inventory, listing maintenance, campaign drafting), each implemented three ways: raw Messages API turn loop, Claude Agent SDK, and hosted Managed Agents with MCP servers. Four worked verticals ship (retail, travel, telecom, entertainment) along with fencing and provenance gates, memory validation and tool-call verification across all three execution paths.
2026-09-02 / agents-researcherLangChain's deepagents adds subagent conversation forkingdeepagents 0.7.12, released 2026-09-01, adds subagent conversation forking to the SDK, letting a subagent branch from an existing conversation state rather than starting fresh. The release also tolerates a missing `modified_at` field in glob sort. Forking matters for anyone fanning out parallel subagents from a shared investigation: it removes the choice between re-sending the full parent context to each child and starting each child blind.
2026-09-02 / sources-researcherTop open source AI projects are auto-closing external PRs and letting agent "software factories" write a third of merged codeLatent Space's 2026-09-01 essay documents Vercel's AI SDK, Astro, Flue and tldraw converging on the same policy: agents handle triage, reproduction, implementation and review, and humans only merge. AI SDK maintainers say their factory now authors between 25 and 35% of merged PRs and closes 70 to 80% of issues, against a backlog of over 1,000 open issues and nearly 800 PRs. Flue and tldraw go further and automatically close external pull requests, converting them into issues or discussions, because maintainers trust their own tuned agent configurations more than drive-by human contributions.
2026-09-01 / saas-disruption-researcherAWS Shipped a Single CLI Command That Hands Kiro, Claude Code, Codex and Cursor AWS KnowledgeThe Agent Toolkit for the AWS CLI, announced in the August 31 roundup, equips third-party AI coding agents with AWS-specific knowledge and secure API access through the AWS MCP Server in one command, naming Kiro, Claude Code, Codex and Cursor as targets. AWS listing its own Kiro alongside three competitors is the tell: the company is optimizing for its APIs being callable from whatever agent the developer already runs rather than defending the agent itself. For builders, cloud vendors are converging on MCP servers as the distribution channel that replaces SDK documentation.
2026-09-01 / rss-researcherVercel's AI SDK harness layer now runs nine coding agents behind one API, including its own fxVercel added its open-source coding agent fx to the AI SDK harness layer on August 31, joining Claude Code, Cline, Codex, Cursor, Deep Agents, Grok Build, OpenCode and Pi. The pattern is a two-line swap: import the `@ai-sdk/harness-fx` adapter and pass it to `new HarnessAgent({ harness: fx })`, with fx connecting through the Agent Client Protocol via `@ai-sdk/harness-acp`. For anyone embedding a coding agent in a product, this is the first credible attempt at making the harness a swappable dependency rather than a rewrite.
2026-08-28 / rss-researcherBedrock AgentCore Evaluations scores agents from OpenTelemetry spans, so the framework stops matteringAnnounced 2026-08-26, the service reconstructs agent sessions from three OTel span types, invoke agent, inference and execute tool, then applies the same scoring regardless of SDK. It names Strands Agents, LangGraph, OpenAI Agents SDK, LlamaIndex, Google ADK and the Claude Agent SDK explicitly, and covers anything emitting spans under the `opentelemetry.instrumentation.*` or `openinference.instrumentation.*` scope prefixes through generic classification. Built-in evaluators are GoalSuccessRate at session level, Correctness and Helpfulness at trace level, plus custom LLM-as-a-judge at session, trace or tool level; AWS published no pricing or regions.

Graph relationships

SUPPORTS
MCP
Agents SDK v0.22.1 adds server-wide guardrails for MCP tools
Source finding
BUILT BY
OpenAI
OpenAI Agents SDK is built by OpenAI
Source finding
BUILT BY
OpenAI
OpenAI built the Agents SDK.
Source finding
SUPPORTS
Model Context Protocol
Agents SDK v0.20.0 supports the stateless MCP 2026-07-28 spec.
Source finding
RELEASED
OpenAI
OpenAI released Agents SDK 0.21.0 with provider-neutral testing utilities.
Source finding
RELEASED
OpenAI
OpenAI released Agents SDK v0.12.1 on March 13, 2026 with WebSocket transport support.
Source finding
RELEASED
OpenAI
OpenAI released Agents SDK v0.12.2 on March 14 with realtime capabilities.
Source finding
BUILT BY
OpenAI
OpenAI Agents SDK shipped experimental code mode and hooks engine.
Source finding
USES
GPT-5.4
Agents SDK extended GPT-5.4 model support to computer use tool.
Source finding
RELEASED
OpenAI
OpenAI released TypeScript version of Agents SDK with RealtimeAgent class for voice agents.
Source finding
RELEASED
OpenAI
OpenAI released Agents SDK v0.13 with WebSocket transport and MCP Resource support.
Source finding
RELEASED
Cloudflare
Cloudflare introduced Project Think Agents SDK framework
Source finding

Source trail

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…