executor
Implementation agent that writes code, runs tests, and commits changes from a concrete plan.
The executor agent is the primary implementation worker in an OMX workflow. It takes a concrete, file-level plan and carries it to a verified, committed result — writing code, running diagnostics, fixing failures, and marking tasks complete. It keeps going until the task is fully resolved rather than stopping at partial progress.
Role
- Implement code changes described in a plan, keeping diffs as small as correctness allows
- Run
lsp_diagnosticson every modified file and resolve all reported errors before finishing - Execute the relevant test suite and fix failures in production code (not in the tests)
- Commit completed work with a meaningful message and update
.omx/state/task records
When invoked
- By
$autopilot,$ralph,$team, and$ralplanonce a plan exists and execution should begin - When a
plannerorralplanoutput lists specific files to touch, tests to add, and a commit message - For targeted fixes when
build-fixerordebuggerhas identified the root cause and a repair strategy - As the worker in
$ultraworkparallel pipelines handling independent task branches
Inputs
- A plan document (from
plannerorralplan) specifying tasks, target files, and expected outcomes - Active
.omx/state/session context indicating which tasks are pending - Optional: LSP diagnostics output, test failure logs, or prior
exploreresults as grounding context
Outputs
- Code changes across the files named in the plan
- Passing test results and clean LSP diagnostics as inline verification evidence
- One or more git commits with descriptive messages
- Updated
.omx/state/entries marking completed tasks
Limits
- Works from a plan — does not improvise system architecture or introduce unrequested abstractions
- Does not skip tests or force-push branches under any circumstances
- Does not claim completion without showing fresh build, diagnostic, or test output as evidence
Related agents
- planner — produces the task plan that
executorconsumes - verifier — independently confirms that the executor's output meets acceptance criteria
- build-fixer — handles stubborn build or type errors that block the executor
- debugger — performs root-cause analysis when executor hits an unexpected failure
Was this page helpful?