docs: add SETUP.MD for Agentic Inner Loop KPI

Adds the SETUP.MD required by the Agentic Inner Loop validation
pipeline (MER KPI 11). Documents prerequisites, environment setup,
build, test, and a prescribed sample code change so an AI agent
can autonomously run the inner dev loop on a clean Linux machine.

PRODEV-623

Author: giuliastf

SETUP.MD

@@ -0,0 +1,91 @@
+# SETUP.md
+
+## Prerequisites
+
+- Python 3.11+
+- [uv](https://docs.astral.sh/uv/) 0.5+
+
+### Supported platforms
+
+This checklist controls the Agentic Inner Loop KPI pipeline targets (clean Linux VM). The repo's own GitHub Actions CI additionally tests on `windows-latest`, but that is not measured by this KPI.
+
+- [x] Linux
+- [ ] Windows
+- [ ] macOS
+
+## Environment Variables
+
+### Standard (injected by pipeline)
+
+None of the standard pipeline variables are required for environment setup, build, or unit tests.
+
+### Project-specific
+
+None. The unit-test suite under the `Test` section below runs fully offline and requires no external authentication. The repo's own GitHub Actions CI passes `UIPATH_URL`, `UIPATH_CLIENT_ID`, and `UIPATH_CLIENT_SECRET` to support optional integration paths, but the default `pytest` invocation does not depend on them and the KPI pipeline does not need to provide them.
+
+## Setup
+
+```bash
+# Install uv if not present
+curl -LsSf https://astral.sh/uv/install.sh | sh
+export PATH="$HOME/.local/bin:$PATH"
+
+# Sync the package with dev dependencies and all optional extras
+uv --directory . sync --all-extras
+```
+
+## Verify Setup
+
+```bash
+python3 --version
+uv --version
+uv --directory . run python -c "import uipath_langchain; print('uipath_langchain ok')"
+```
+
+## Build
+
+```bash
+uv --directory . build
+```
+
+## Test
+
+```bash
+uv --directory . run pytest
+```
+
+## Sample Code Change
+
+### The change
+
+Add a new `get_execution_job_id` helper to `src/uipath_langchain/_utils/_environment.py`, immediately after the existing `get_default_timeout` function. The helper is the symmetric complement of the existing `get_execution_folder_path` — it reads the `UIPATH_JOB_KEY` env var at runtime:
+
+```python
+def get_execution_job_id() -> str | None:
+    """Reads the agent's executing job key from the runtime environment."""
+    return os.environ.get("UIPATH_JOB_KEY")
+```
+
+Then create `tests/utils/test_environment_job_id.py` with two pytest tests:
+
+```python
+import pytest
+
+from uipath_langchain._utils._environment import get_execution_job_id
+
+
+def test_get_execution_job_id_unset(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.delenv("UIPATH_JOB_KEY", raising=False)
+    assert get_execution_job_id() is None
+
+
+def test_get_execution_job_id_set(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.setenv("UIPATH_JOB_KEY", "job-abc-123")
+    assert get_execution_job_id() == "job-abc-123"
+```
+
+### Verification
+
+```bash
+uv --directory . run pytest tests/utils/test_environment_job_id.py -v
+```