agent-os and agent-sre

These are ecosystem siblings—both deprecated projects that were consolidated into a single unified governance toolkit, suggesting they previously addressed different aspects (OS-level agent management and SRE operations) of the same agent governance problem space.

agent-os
38
Emerging
agent-sre
26
Experimental
Maintenance 0/25
Adoption 8/25
Maturity 11/25
Community 19/25
Maintenance 0/25
Adoption 4/25
Maturity 9/25
Community 13/25
Stars: 67
Forks: 20
Downloads:
Commits (30d): 0
Language: Python
License: MIT
Stars: 7
Forks: 2
Downloads:
Commits (30d): 0
Language: Python
License: MIT
Archived No Package No Dependents
Archived No Package No Dependents

About agent-os

imran-siddique/agent-os

[DEPRECATED] Moved to microsoft/agent-governance-toolkit

Provides a kernel-based governance layer with policy enforcement, sandbox isolation, and audit logging for autonomous agents across multiple LLM frameworks (LangGraph, OpenAI Agents, Dify, LlamaIndex). Implements Model Context Protocol (MCP) integration for extensible tool management and covers 9 of 10 OWASP Agentic Top 10 vulnerabilities. Achieves sub-0.1ms p99 latency overhead while maintaining framework-agnostic compatibility through adapter patterns.

About agent-sre

imran-siddique/agent-sre

[DEPRECATED] Moved to microsoft/agent-governance-toolkit

Agent SRE provides reliability engineering for AI agent systems, implementing SLOs, error budgets, chaos testing, progressive delivery, and cost guardrails. It integrates directly into agent frameworks like LangChain, LlamaIndex, Dify, and AutoGen, leveraging OpenTelemetry for native OTLP export to platforms such as Langfuse, LangSmith, and Prometheus. The architecture includes dedicated engines for SLOs, error budgets, chaos injection, circuit breaking, canary deployments, and cost management to address common AI agent reliability challenges.

Related comparisons

Scores updated daily from GitHub, PyPI, and npm data. How scores work