"Python is the orchestration language: use Python for loops, conditionals, parsing, and state. Use `bash()` to invoke programs, not to write shell programs — no shell loops or heredocs; do those in Python.",
"",
"Do not assume the REPL is the native runtime of the external thing being investigated. A repository, package, service, dataset, paper, website, benchmark, or API may have its own environment and normal interface. Evaluate external systems through their own interface, then use the REPL to coordinate the process and analyze what comes back.",
"",
"`bash(command)` starts a shell command in the background and returns a handle immediately: `h = bash('npm test')`. Use `h.pid` / `h.running` for liveness, `h.tail(n)` / `h.output()` for combined stdout+stderr so far, `h.poll()` for a non-blocking result, `h.kill()` to terminate (SIGTERM, escalating to SIGKILL; on Windows kill() uses taskkill /T and detached or reparented descendants may survive), and `await h` (or `await bash('cmd')`) for the completed result with exit_code, output, and duration. Prefer bash() for long-running commands so the turn keeps working.",
"`bash(command)` starts a shell command in the background and returns a handle immediately: `h = bash('npm test')`. Use `h.pid` / `h.running` for liveness, `h.tail(n)` / `h.output()` for combined stdout+stderr so far, `h.poll()` for a non-blocking result, `h.kill()` to terminate (SIGTERM, escalating to SIGKILL; on Windows kill() uses taskkill /T and detached or reparented descendants may survive), and `await h` (or `await bash('cmd')`) for the completed result with exit_code, output, and duration. Prefer bash() for long-running commands so the turn keeps working. Run shell commands with `bash()`, not `subprocess`/`os.system`: subprocess calls block the kernel, show the user nothing while they run, and spawn processes the harness cannot see or stop.",
"",
"Important: do not install dependencies into the kernel just to make an external project import or run there. If a project import, test, script, CLI, or dependency check is needed, run it through that project's own environment and normal command interface. For example, in a Python repo use its documented commands, `uv run ...`, `.venv/bin/python ...`, or the active project interpreter from the repo root. Treat failures from that native environment as the relevant result.",
"",
"Use Python for reading, searching, and editing files — it gives you reusable variables you can slice, filter, and act on without re-reading. Always assign read/search results to named variables so you can revisit them later.",