Dagger for Entire
Dagger can now run your checks in Entire, a new Git forge for agent-human collaboration. Entire replaces pull requests with “Trails”, which bundle code and agent sessions. Dagger lets you run the same pipelines before and after push, with shared caching and scalable compute.
Try it now
Set up Dagger for Entire: 1. If needed, set up Dagger: https://docs.dagger.io/getting-started 2. Run `dagger cloud checks on`, so that Dagger Cloud runs my checks on every push. 3. Walk me through a test Trail, to verify that it all works.
Agent-native CI for an agent-native forge
Last week, Entire launched a new Git forge, for the same reason we launched a new CI platform: coding agents broke the software delivery stack. Superficial upgrades won’t fix it; it must be rebuilt from the ground up.
Entire rebuilt the forge, and we rebuilt CI: it’s a perfect match. So we got our teams together and built a native integration: Dagger for Entire.
Set up Dagger locally, run dagger cloud checks on, follow the instructions, and you’re done. Every push triggers the same Dagger checks you run locally, executed and cached by Dagger Cloud.
Goodbye pull requests, hello Trails
Entire replaces pull requests with a new concept: Trails. Trails store the code and the agent sessions together, so that you can review the code with its full context. Each time your agent pushes, Dagger runs your checks again and annotates the Trail with the results.
Before and after push
Your agent can run the same checks before push, in its local context, with automatic caching.
By default, checks run on your machine. To support many concurrent sessions without melting your machine, enable remote execution: export DAGGER_ENGINE=cloud and continue using dagger as usual. It now transparently hands off pipeline execution to Dagger Cloud, streaming local files, secrets and network tunnels as needed. Execution is sandboxed, so your agents and CI can safely share one cache: each operation runs only once, before or after push, including the builds that your checks depend on.
All the context your agent needs to debug
When a check fails, your agent does not need to parse noisy logs. Each check links to a high-resolution trace, with the full context of the failure, so the agent can debug accurately without wasting tokens.
The agent can also run the same checks again from its terminal, against any branch or tag of the Entire repository. dagger shell opens a shell in the intermediate artifacts of your checks, such as a build’s output or the container that your tests run in. With remote execution, the shell runs on Dagger Cloud and is streamed to your local terminal.
Join us on Discord and tell us what you think.
Set up Dagger for Entire: 1. If needed, set up Dagger: https://docs.dagger.io/getting-started 2. Run `dagger cloud checks on`, so that Dagger Cloud runs my checks on every push. 3. Walk me through a test Trail, to verify that it all works.