Ross ROSS = Recommend OSS · open-source software intelligence for agents

omnigent-ai/omnigent

Omnigent is an open-source AI agent framework and meta-harness: orchestrate Claude Code, Codex, Cursor, Pi, and custom agents — swap harnesses without rewriting, enforce policies and sandboxing, and collaborate in real time from any device. observed · 2026-08-28

github.com/omnigent-ai/omnigent · homepage · Python · Apache-2.0 (permissive) observed · 2026-08-28

Health v2 · maintenance only

80/100

  • Activity 99
  • Release rhythm 99
  • Longevity 5

Flags: young

How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.

  • gap_med: 6.0
  • age_days: 83
  • days_rel: 8
  • days_push: 7
  • n_releases_24m: 13

Full methodology

Adoption not part of the score

9314 stars · 1432 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Omnigent is an open-source meta-harness that provides a common orchestration layer over AI coding agents like Claude Code, Codex, Cursor, and custom YAML-defined agents. It adds sandboxing, stateful contextual policies, cloud sandbox hosting, and real-time multi-device collaboration on top of any agent harness.

Use cases

  • run claude code and codex through one unified interface
  • supervise multiple ai coding agents in the same session
  • enforce spend caps and approval gates on ai agents
  • run coding agents in cloud sandboxes instead of my laptop
  • check on long-running agent tasks from my phone
  • share a live agent session with teammates for pair programming
  • run agents unattended without exposing my credentials
  • build custom multi-agent orchestrators in yaml

When to choose

  • you use multiple agent harnesses (Claude Code, Codex, Cursor, etc.) and want one orchestration, policy, and collaboration layer over all of them
  • you need stateful guardrails like cost budgets, risk-based approval escalation, or model routing that static tool rules can't express
  • you want to run agents unattended in OS-level or cloud sandboxes with filesystem, network, and credential isolation
  • you want to start agent sessions in a terminal and continue them from a browser, phone, or desktop app with real-time sync

When to avoid

  • you need a battle-tested production tool - Omnigent is explicitly alpha with a pre-1.0 API
  • you use a single agent harness and are happy with its native terminal UI and workflow
  • you want a lightweight library to embed agent logic in your own code rather than a full platform with server, CLI, and apps
  • you need Windows-native sandboxing - the OS sandbox relies on bubblewrap/seccomp (Linux) and Seatbelt (macOS)

Facets

framework · maturity experimental

agent-framework security middleware web-framework cli chat-interface large-language-models developer-tools security self-hosted python windows browser cli self-hosted cross-platform meta-harness agent-orchestration coding-agents multi-agent policies sandboxing cloud-sandboxes real-time-collaboration cost-control credential-brokering yaml-agents harness-agnostic sandbox ai-agents automation macos linux ios android docker kubernetes

10 sources

Member repositories

RepositoryRoleHealth v2
omnigent-ai/omnigentmain80

For agents

markdown · JSON · MCP: product_card(name="omnigent-ai/omnigent")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem