Fetching from the wire…
Tools2026-09-23 · source-backed
PR #8886, in 1.2.12, accepts a JSON Schema dict, Pydantic model, TypedDict or dataclass. Pydantic-style types validate on resume and an invalid value raises from the node before it commits. Studio renders a typed form instead of a free-form JSON box. That closes the gap where a malformed approval payload flowed straight into downstream state. Default is None, so existing graphs are unaffected.
Each link below shares sources, entities, or timing with this story.
FastMCP v4.0.1 (September 2) makes ClientGroup reference-count its context like Client does, so entering a connected group from a nested block reuses connections instead of raising (release). LangChain's 1.4.0a4, cut hours later, contains commits adding _ReentrantClientGroup,...
semantica-agi/semantica jumped to 3,677 stars. MIT, pip install semantica, v0.6.0, 2,229 commits. The pipeline is fully deterministic, Ingest → Parse → Normalize → Split → Extract → Conflict Detection → Deduplication → Knowledge Graph → Ontology/Reasoning/Provenance/Decisions,...
3,364 stars since its August 17 creation. Every action against a computer, file, MCP server or UI component routes through a single gateway that resolves the target, decides it against policy, writes an audit row, then acts or refuses while naming the rule. Each bot gets its o...
LangGraph 1.1 ships opt-in version="v2" streaming that yields strongly-typed StreamPart dicts with type, namespace, data, and interrupts — eliminating the untyped dict problem that plagued production observability. Pydantic model coercion is automatic. Time-travel with interru...
MIT-licensed, built on LangChain and LangGraph, v2.0 expanding from research-only to sandboxed filesystem, persistent memory, a skills system, and sub-agent orchestration with Slack/Telegram/Feishu integrations (GitHub). The pattern across every framework today is identical: m...
AgentBouncr sits between AI agents and their tools using deterministic (not LLM-based) policy enforcement. JSON policy engine with 11 condition operators, SHA-256 hash-chained audit trails, synchronous kill switch, and injection detection. Works with LangChain, Vercel AI SDK,...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.