How to batch 5 coding-agent turns into 1 macro workflow

#ai #opensource #testing #rag

Disclosure: I maintain Tura.

Most coding agents handle a routine change like this:

Normal tool-calling agent

Turn 1 — inspect

rg -n "TODO|command_run|handler" crates/
rg --files crates/runtime/src crates/tools/src

Turn 2 — apply the patch

Turn 3 — build

cargo build -p runtime

Turn 4 — test

cargo test -p runtime --lib

Turn 5 — lint

cargo clippy -p runtime --all-targets

The model wakes up five times, even though most of the sequence is predictable.

Tura macro workflow

Tura exposes one macro tool called command_run. The agent can send the same work once:

{
  "name": "command_run",
  "arguments": {
    "commands": [
      { "step": 1, "command_type": "shell_command", "command_line": "rg -n \"TODO|command_run|handler\" crates/" },
      { "step": 1, "command_type": "shell_command", "command_line": "rg --files crates/runtime/src crates/tools/src" },
      { "step": 2, "command_type": "apply_patch", "command_line": "apply the patch" },
      { "step": 3, "command_type": "shell_command", "command_line": "cargo build -p runtime" },
      { "step": 4, "command_type": "shell_command", "command_line": "cargo test -p runtime --lib" },
      { "step": 4, "command_type": "shell_command", "command_line": "cargo clippy -p runtime --all-targets" }
    ]
  }
}

The build, tests, and lint still run. The difference is that the model does not need a new turn between every predictable step. If something unexpected happens, control returns to the model.

In the published DeepSWE comparison, Tura Direct used 69.1% fewer turns and 77.5% fewer tokens than Codex CLI. Tura Balanced used 35.8% fewer turns and 31.1% fewer tokens while reaching an 80% success rate.