Подборка — 3 881

Architecture pattern for autonomous agents using Claude Cowork, Notion, and MCP. Persistent multi-domain agents with no custom infrastructure or code.

AI outbound CLI. Source 450M+ contacts, email + LinkedIn campaigns, conversations. Agent-native JSON output.

Chat with your data. NL to Pandas/SQL.

Intent signal CLI. LinkedIn engagers, keyword posters, job changers, funding events. JSON output for agent pipelines.

Live data RAG. Real-time streaming. 50k+ stars.

Organizational context and memory for AI agents. 60-tool MCP server, 100+ integrations.

OSS ChatGPT alternative. 100% offline.

LLMs as single files. Zero setup. Mozilla.

Fork-and-run template for 24/7 autonomous AI agents. Pre-configured SOUL.md, memory system, KANBAN, heartbeat. Start in 30 minutes.

Self-hosted AI agent platform. Persistent memory (hybrid search + LLM re-ranking), 23+ providers (including Ollama), plugin store, mini-apps SDK, cron scheduling, 6 messaging channels. SQLite, runs on a Pi.

Open-source multi-agent orchestration with web dashboard, task lifecycle, knowledge base, and autopilot mode. Built on Claude Code. Runs locally.

4-role AI team (PM+DEV+OPS+QA) in Cursor IDE. File-based task routing, auto patrol bot. 87 person-days in 17 days.

Industry-standard API spec. Foundation for agent tools.

CLI to transform OpenAPI specs into 9 agent formats (MCP, AGENTS.md, Claude tools, etc.). npx agentify-cli .

Auto-rollback for AI agent config changes. Monitors health endpoint, reverts config + restarts service on failures. Zero deps.

Industry standard for coding agents. Top: 80.9% (Claude Opus 4.6).

8-environment LLM agent benchmark.

Web agent benchmark. Real websites.

Google

Zhipu

Microsoft

Structural, type, quality guarantees for LLM outputs.

Security toolkit. Input/output scanning.

Prompt injection detection.

AI pentesting, vuln discovery, red teaming. HITL.

Graph-based cybersecurity simulation.

GPT-powered pentesting. Automated reasoning.

OSS security scanner for AI agent skills/plugins/MCP servers. Pre-install taint tracking, post-uninstall residue detection.

AI-assisted reminiscence therapy for elderly cognitive training. Narrative quality scoring (v0.7), life story book generation. Chinese + English.