omnigent

omnigent is an open-source AI developer tool for builders evaluating agentic workflows, model-development infrastructure, or coding-agent operations. The canonical source is the public GitHub repository at https://github.com/omnigent-ai/omnigent. OpenTools verified that repository during this creation pass and used public metadata rather than marketing copy: 2,918 stars, 335 forks, 146 open issues, and a latest public push dated 2026-06-17. The project description is: A meta-harness for all your AI agents. Omnigent provides a common layer over Claude Code, Codex, Pi, and the agents you write yourself: swap or combine harnesses without rewriting, keep them in check with policies and sandboxing, and collaborate in real time on the same live session, from any device.. Omnigent is a meta-harness for agent work across tools such as Claude Code and Codex, which makes policy, sandboxing, shared sessions, and orchestration the core evaluation points. That framing matters. Many AI projects look similar in a queue, but this listing is for a runnable or inspectable developer tool, not a foundation model and not a generic article. The practical question for builders is whether it makes agent or model work easier to run, review, repeat, or govern. A good first test is simple. Read the upstream README, clone the project in a disposable environment, and run only the documented setup path. Check exactly which files it reads, which commands it can execute, what credentials it expects, and whether it calls external model APIs. If the project integrates with Claude Code, Codex, local models, or other agents, test it against a throwaway repository before giving it production code. omnigent should interest developers, AI engineering teams, and platform teams that already work with coding agents or foundation-model infrastructure. Solo builders can use it to experiment with a more structured workflow. Larger teams should evaluate it as part of an internal developer-tooling stack, with the same security review they would apply to any automation that can change code, run commands, or broker model calls. Pricing is listed as free/open-source access because the source repository is public. That does not mean every workflow is free. Users may still pay for connected LLM APIs, hosted compute, cloud runners, storage, or private services they attach to the project. Review the Apache License 2.0 terms, recent commits, issue activity, and dependency tree before adopting it commercially. OpenTools keeps this page source-backed and conservative. We verified the repository URL, public GitHub activity, description, and project positioning. We did not infer hidden benchmarks, private roadmap details, or paid plans that are not documented upstream. If the project later adds hosted pricing, docs, screenshots, or a stable release channel, this record should be updated with those facts instead of creating a second duplicate listing.