hi, this is agentconnect. Sandboxes are seen as the standard way to keep agents safe. But problems arise when team runs several agents with different levels of access. That's why we want to solve.
Access to every repo, mcp server, skill, and credential is controlled separately. A2A permissions are set separately in each direction. Delegated tasks run on receiving agent’s side. Only the result comes back. Access is never passed along with it. This keeps permissions isolated and secure.
We have tested it with claude code, codex, and deepseek harness. You can run it with docker compose or deploy it on kubernetes with agent sandbox. Happy to answer questions.
Main difference is agentconnect’s team collaboration layer.
1. Public facing agents can answer support questions. But they cannot access resources shared with private agents.
2. DM between a team member and an agent can stay private. Other team members cannot see them.
3. Each agent can have its own slack or telegram bot. Multiple agents can also work together in the same channel or thread. In hermes, each profile has its own messaging gateway and bot setup.
4. Each agent can access different code repos. It can also respond directly to issues, pull requests, code reviews, and check results.
Another difference is the agent framework. Agentconnect uses acp. This lets teams run native codex, claude code, and other compatible agents. Hermes uses its own framework and supports models from different providers.
In my experience, native coding agents are often easier to use for engineering teams. They also tend to produce better coding results. However, they may use more tokens than hermes.
6 comments
[ 3.4 ms ] story [ 28.9 ms ] threadAccess to every repo, mcp server, skill, and credential is controlled separately. A2A permissions are set separately in each direction. Delegated tasks run on receiving agent’s side. Only the result comes back. Access is never passed along with it. This keeps permissions isolated and secure.
We have tested it with claude code, codex, and deepseek harness. You can run it with docker compose or deploy it on kubernetes with agent sandbox. Happy to answer questions.