lcode 0.18.0

lcode without a person at the keyboard (#48).

• Structured output.

  • lcode -p "…" --output json prints one result object: status, answer, tool calls, files changed, token usage.
  • --output stream-json prints an event per model step and tool result, then the result.
  • --max-steps and --allowed-tools limit a run, and the exit code says how it ended (0 success, 1 error, 3 max steps). Anything that needs permission is refused, since nobody can be asked.

• A Python API.

  from lcode import Session
  
  with Session("path/to/repo", permission_mode="auto-edit") as session:
      result = session.run("Fix the failing test")
      print(result.status, result.files_changed)

  stream() yields the events, and approve decides permission requests. See examples/api/fix_tests.py.

• A GitHub Action for self-hosted runners. uses: nasser1941/lcode@v0.18.0 on a machine of yours with a GPU:

  • @lcode comments: it answers them on issues and pull requests, from the repository's owners, members and collaborators only.
  • Changes: it proposes them as a pull request that closes the issue, or pushes them to the pull request's branch.
  • Reviews: it can review new pull requests.
  • Privacy: the code and the model stay on your runner.

Docs: Automation, with the JSON schema, the API and an example workflow.

Full changelog: v0.17.0...v0.18.0