diff --git a/meshtastic/tests/test_inject_nanopb_options.py b/meshtastic/tests/test_inject_nanopb_options.py
new file mode 100644
index 0000000..2d89188
--- /dev/null
+++ b/meshtastic/tests/test_inject_nanopb_options.py
@@ -0,0 +1,605 @@
+"""Tests for bin/inject_nanopb_options.py — the nanopb options injection script
+and the generated protobuf descriptors it produces.
+
+Part 1 (test_parse_*, test_inject_*): unit-tests the script's logic directly,
+using small synthetic proto snippets.
+
+Part 2 (test_descriptor_*): smoke-tests the already-generated _pb2.py files to
+confirm the regen pipeline embedded the expected nanopb options.
+"""
+
+import importlib.util
+import sys
+import textwrap
+from pathlib import Path
+from typing import Any, Dict, Tuple
+from unittest.mock import patch
+
+import pytest
+
+# ---------------------------------------------------------------------------
+# Load bin/inject_nanopb_options.py as a module without adding it to the
+# package.  __main__ guard means no side-effects on import.
+# ---------------------------------------------------------------------------
+_SCRIPT_PATH = Path(__file__).parent.parent.parent / "bin" / "inject_nanopb_options.py"
+
+
+def _load_inject_module():
+    spec = importlib.util.spec_from_file_location("inject_nanopb_options", _SCRIPT_PATH)
+    mod = importlib.util.module_from_spec(spec)
+    with patch.object(sys, "argv", ["inject_nanopb_options.py"]):
+        spec.loader.exec_module(mod)
+    return mod
+
+
+_inj = _load_inject_module()
+
+parse_value = _inj.parse_value
+parse_options_file = _inj.parse_options_file
+format_nanopb_opts = _inj.format_nanopb_opts
+inject_into_proto = _inj.inject_into_proto
+message_path_matches = _inj.message_path_matches
+
+# Convenience: the nanopb import path the script uses after the sed fixup
+NANOPB_IMPORT = 'import "meshtastic/protobuf/nanopb.proto";'
+
+
+# ===========================================================================
+# Part 1 — Script unit tests
+# ===========================================================================
+
+# ---------------------------------------------------------------------------
+# parse_value
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.unit
+def test_parse_value_integer():
+    assert parse_value("40") == 40
+
+
+@pytest.mark.unit
+def test_parse_value_negative_integer():
+    assert parse_value("-1") == -1
+
+
+@pytest.mark.unit
+def test_parse_value_true():
+    assert parse_value("true") is True
+
+
+@pytest.mark.unit
+def test_parse_value_false():
+    assert parse_value("false") is False
+
+
+@pytest.mark.unit
+def test_parse_value_string():
+    assert parse_value("IS_8") == "IS_8"
+
+
+# ---------------------------------------------------------------------------
+# parse_options_file
+# ---------------------------------------------------------------------------
+
+
+def _write_options(tmp_path: Path, content: str) -> Path:
+    p = tmp_path / "test.options"
+    p.write_text(textwrap.dedent(content))
+    return p
+
+
+@pytest.mark.unit
+def test_parse_wildcard(tmp_path):
+    """Wildcard pattern (no dot) lands in the wildcard dict."""
+    f = _write_options(tmp_path, "*macaddr max_size:6 fixed_length:true\n")
+    specific, wildcard = parse_options_file(f)
+    assert "macaddr" in wildcard
+    assert wildcard["macaddr"] == {"max_size": 6, "fixed_length": True}
+    assert specific == {}
+
+
+@pytest.mark.unit
+def test_parse_specific(tmp_path):
+    """Single-dot pattern lands in the specific dict with a 2-tuple key."""
+    f = _write_options(tmp_path, "*User.long_name max_size:40\n")
+    specific, wildcard = parse_options_file(f)
+    assert ("User", "long_name") in specific
+    assert specific[("User", "long_name")] == {"max_size": 40}
+    assert wildcard == {}
+
+
+@pytest.mark.unit
+def test_parse_multilevel(tmp_path):
+    """Three-part pattern (Route.Link.uid) produces a 3-tuple key."""
+    f = _write_options(tmp_path, "*Route.Link.uid max_size:48\n")
+    specific, wildcard = parse_options_file(f)
+    assert ("Route", "Link", "uid") in specific
+    assert specific[("Route", "Link", "uid")] == {"max_size": 48}
+
+
+@pytest.mark.unit
+def test_parse_strips_inline_comments(tmp_path):
+    """Text after # is ignored."""
+    f = _write_options(tmp_path, "*id max_size:16 # node id strings\n")
+    _, wildcard = parse_options_file(f)
+    assert wildcard["id"] == {"max_size": 16}
+
+
+@pytest.mark.unit
+def test_parse_skips_comment_only_lines(tmp_path):
+    """Lines that are entirely comments produce no entries."""
+    f = _write_options(tmp_path, "# this is a comment\n*id max_size:16\n")
+    _, wildcard = parse_options_file(f)
+    assert list(wildcard.keys()) == ["id"]
+
+
+@pytest.mark.unit
+def test_parse_skips_blank_lines(tmp_path):
+    """Blank lines are silently ignored."""
+    f = _write_options(tmp_path, "\n\n*id max_size:16\n\n")
+    _, wildcard = parse_options_file(f)
+    assert "id" in wildcard
+
+
+@pytest.mark.unit
+def test_parse_skips_non_python_options(tmp_path):
+    """Options not in FIELD_OPTIONS (e.g. anonymous_oneof) are dropped."""
+    f = _write_options(tmp_path, "*MeshPacket.payload_variant anonymous_oneof:true\n")
+    specific, wildcard = parse_options_file(f)
+    # anonymous_oneof is not in FIELD_OPTIONS → no entry should be produced
+    assert specific == {}
+    assert wildcard == {}
+
+
+@pytest.mark.unit
+def test_parse_merges_repeated_patterns(tmp_path):
+    """Two lines for the same pattern are merged."""
+    f = _write_options(
+        tmp_path,
+        "*SecurityConfig.admin_key max_size:32\n"
+        "*SecurityConfig.admin_key max_count:3\n",
+    )
+    specific, _ = parse_options_file(f)
+    assert specific[("SecurityConfig", "admin_key")] == {"max_size": 32, "max_count": 3}
+
+
+@pytest.mark.unit
+def test_parse_int_and_bool_values(tmp_path):
+    """int_size parses as int; fixed_length parses as bool."""
+    f = _write_options(tmp_path, "*Data.payload max_size:233 fixed_length:false\n")
+    specific, _ = parse_options_file(f)
+    opts = specific[("Data", "payload")]
+    assert opts["max_size"] == 233
+    assert opts["fixed_length"] is False
+
+
+# ---------------------------------------------------------------------------
+# message_path_matches
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.unit
+def test_message_path_matches_simple():
+    stack = [("message", "User")]
+    assert message_path_matches(stack, ("User",))
+
+
+@pytest.mark.unit
+def test_message_path_matches_nested():
+    stack = [("message", "Config"), ("message", "DeviceConfig")]
+    assert message_path_matches(stack, ("DeviceConfig",))
+    assert message_path_matches(stack, ("Config", "DeviceConfig"))
+
+
+@pytest.mark.unit
+def test_message_path_matches_with_oneof_in_stack():
+    """oneof frames in the stack are skipped when looking for messages."""
+    stack = [("message", "MeshPacket"), ("oneof", "payload_variant")]
+    assert message_path_matches(stack, ("MeshPacket",))
+
+
+@pytest.mark.unit
+def test_message_path_no_match():
+    stack = [("message", "User")]
+    assert not message_path_matches(stack, ("Route",))
+
+
+@pytest.mark.unit
+def test_message_path_multilevel_partial_match():
+    """A 2-element path must match the last 2 message names on the stack."""
+    stack = [("message", "Route"), ("message", "Link")]
+    assert message_path_matches(stack, ("Route", "Link"))
+    assert not message_path_matches(stack, ("Other", "Link"))
+
+
+# ---------------------------------------------------------------------------
+# format_nanopb_opts
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.unit
+def test_format_max_size():
+    assert format_nanopb_opts({"max_size": 40}) == "(nanopb).max_size = 40"
+
+
+@pytest.mark.unit
+def test_format_int_size_as_enum():
+    assert format_nanopb_opts({"int_size": 8}) == "(nanopb).int_size = IS_8"
+    assert format_nanopb_opts({"int_size": 16}) == "(nanopb).int_size = IS_16"
+    assert format_nanopb_opts({"int_size": 32}) == "(nanopb).int_size = IS_32"
+    assert format_nanopb_opts({"int_size": 64}) == "(nanopb).int_size = IS_64"
+
+
+@pytest.mark.unit
+def test_format_bool_true():
+    assert format_nanopb_opts({"fixed_length": True}) == "(nanopb).fixed_length = true"
+
+
+@pytest.mark.unit
+def test_format_bool_false():
+    assert format_nanopb_opts({"fixed_length": False}) == "(nanopb).fixed_length = false"
+
+
+# ---------------------------------------------------------------------------
+# inject_into_proto — helpers
+# ---------------------------------------------------------------------------
+
+_NANOPB_IMPORT_PATH = "meshtastic/protobuf/nanopb.proto"
+
+
+def _inject(proto_src: str, specific=None, wildcard=None) -> str:
+    """Run inject_into_proto with empty dicts as defaults."""
+    return inject_into_proto(
+        textwrap.dedent(proto_src),
+        specific or {},
+        wildcard or {},
+        _NANOPB_IMPORT_PATH,
+    )
+
+
+# ---------------------------------------------------------------------------
+# inject_into_proto — option injection
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.unit
+def test_inject_adds_option_to_plain_field():
+    """A field with no existing options gets a nanopb annotation."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/channel.proto";
+        message User {
+          string long_name = 1;
+        }
+    """
+    result = _inject(proto, specific={("User", "long_name"): {"max_size": 40}})
+    assert "long_name = 1 [(nanopb).max_size = 40];" in result
+
+
+@pytest.mark.unit
+def test_inject_merges_with_existing_options():
+    """nanopb annotation is appended after existing field options."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/channel.proto";
+        message User {
+          bytes macaddr = 4 [deprecated = true];
+        }
+    """
+    result = _inject(proto, wildcard={"macaddr": {"max_size": 6}})
+    assert "[deprecated = true, (nanopb).max_size = 6];" in result
+
+
+@pytest.mark.unit
+def test_inject_int_size_uses_enum_name():
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          uint32 hop_limit = 9;
+        }
+    """
+    result = _inject(proto, specific={("Foo", "hop_limit"): {"int_size": 8}})
+    assert "(nanopb).int_size = IS_8" in result
+
+
+@pytest.mark.unit
+def test_inject_wildcard_applied_across_messages():
+    """A wildcard option hits the matching field in every message."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message A {
+          bytes macaddr = 1;
+        }
+        message B {
+          bytes macaddr = 2;
+        }
+    """
+    result = _inject(proto, wildcard={"macaddr": {"max_size": 6}})
+    assert result.count("(nanopb).max_size = 6") == 2
+
+
+@pytest.mark.unit
+def test_inject_specific_not_leaking_to_other_messages():
+    """A message-specific option does NOT apply to a different message's field."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message User {
+          string id = 1;
+        }
+        message Other {
+          string id = 1;
+        }
+    """
+    result = _inject(proto, specific={("User", "id"): {"max_size": 16}})
+    lines = result.splitlines()
+    user_line = next(l for l in lines if "User" not in l and "id = 1" in l and "Other" not in l.split("message")[0] if "message" not in l)
+    # Easier: count annotations — should be exactly one
+    assert result.count("(nanopb).max_size = 16") == 1
+
+
+@pytest.mark.unit
+def test_inject_nested_message():
+    """A 2-part specific key only hits the field in the correct nested message."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Route {
+          message Link {
+            string uid = 1;
+          }
+          string uid = 2;
+        }
+    """
+    # Route.Link.uid → key = ('Route', 'Link', 'uid')
+    result = _inject(proto, specific={("Route", "Link", "uid"): {"max_size": 48}})
+    lines = result.splitlines()
+    # Only the uid inside Link should have the annotation
+    assert result.count("(nanopb).max_size = 48") == 1
+    # Confirm it's the inner one (it has 4 spaces more indent than outer uid)
+    annotated = next(l for l in lines if "(nanopb).max_size = 48" in l)
+    plain = next(l for l in lines if "uid = 2" in l)
+    assert annotated.index("uid") > plain.index("uid")
+
+
+@pytest.mark.unit
+def test_inject_skips_enum_body_values():
+    """Enum value lines must not be treated as field declarations."""
+    proto = """\
+        syntax = "proto3";
+        message Foo {
+          enum Role {
+            CLIENT = 0;
+            ROUTER = 2;
+          }
+          Role role = 1;
+        }
+    """
+    # Wildcard for 'role' should only hit the field, not enum values
+    result = _inject(proto, wildcard={"role": {"max_size": 8}})
+    assert result.count("(nanopb)") == 1
+    assert "(nanopb)" not in next(l for l in result.splitlines() if "CLIENT" in l)
+
+
+@pytest.mark.unit
+def test_inject_optional_qualifier_preserved():
+    """The 'optional' qualifier is kept when a field gets an annotation."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          optional uint32 altitude = 3;
+        }
+    """
+    result = _inject(proto, specific={("Foo", "altitude"): {"int_size": 16}})
+    assert "optional uint32 altitude = 3 [(nanopb).int_size = IS_16];" in result
+
+
+@pytest.mark.unit
+def test_inject_repeated_qualifier_preserved():
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          repeated int32 snr = 2;
+        }
+    """
+    result = _inject(proto, specific={("Foo", "snr"): {"max_count": 8}})
+    assert "repeated int32 snr = 2 [(nanopb).max_count = 8];" in result
+
+
+@pytest.mark.unit
+def test_inject_multiple_options_on_one_field():
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          repeated int32 snr = 1;
+        }
+    """
+    result = _inject(proto, specific={("Foo", "snr"): {"max_count": 8, "int_size": 8}})
+    assert "(nanopb).max_count = 8" in result
+    assert "(nanopb).int_size = IS_8" in result
+
+
+# ---------------------------------------------------------------------------
+# inject_into_proto — import insertion
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.unit
+def test_inject_adds_nanopb_import_when_absent():
+    """nanopb.proto import is added when the file has other imports."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          string name = 1;
+        }
+    """
+    result = _inject(proto, wildcard={"name": {"max_size": 30}})
+    assert NANOPB_IMPORT in result
+
+
+@pytest.mark.unit
+def test_inject_no_duplicate_nanopb_import():
+    """nanopb.proto import is NOT added a second time if already present."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        import "meshtastic/protobuf/nanopb.proto";
+        message Foo {
+          string name = 1;
+        }
+    """
+    result = _inject(proto, wildcard={"name": {"max_size": 30}})
+    assert result.count(NANOPB_IMPORT) == 1
+
+
+@pytest.mark.unit
+def test_inject_import_placed_after_existing_imports():
+    """nanopb import appears after the last existing import, not at the top."""
+    proto = """\
+        syntax = "proto3";
+        import "meshtastic/protobuf/mesh.proto";
+        message Foo {
+          string name = 1;
+        }
+    """
+    result = _inject(proto, wildcard={"name": {"max_size": 30}})
+    lines = result.splitlines()
+    mesh_idx = next(i for i, l in enumerate(lines) if "mesh.proto" in l)
+    nanopb_idx = next(i for i, l in enumerate(lines) if "nanopb.proto" in l)
+    assert nanopb_idx == mesh_idx + 1
+
+
+@pytest.mark.unit
+def test_inject_import_after_syntax_when_no_existing_imports():
+    """When a proto has no imports, nanopb import goes AFTER the syntax line,
+    not before it (regression test for the last_import_idx == -1 bug)."""
+    proto = """\
+        syntax = "proto3";
+        message XModem {
+          uint32 seq = 2;
+        }
+    """
+    result = _inject(proto, specific={("XModem", "seq"): {"int_size": 16}})
+    lines = result.splitlines()
+    syntax_idx = next(i for i, l in enumerate(lines) if l.strip().startswith("syntax"))
+    nanopb_idx = next(i for i, l in enumerate(lines) if "nanopb.proto" in l)
+    assert nanopb_idx > syntax_idx, "nanopb import must come after the syntax line"
+    # syntax line must still be first non-blank line
+    first_non_blank = next(l.strip() for l in lines if l.strip())
+    assert first_non_blank.startswith("syntax")
+
+
+@pytest.mark.unit
+def test_inject_noop_when_no_options():
+    """Proto file is returned unchanged when there are no options to inject."""
+    proto = 'syntax = "proto3";\nmessage Foo { string x = 1; }\n'
+    result = _inject(proto)
+    assert result == proto
+
+
+# ===========================================================================
+# Part 2 — Descriptor integration tests
+# Verify that regen-protobufs.sh produced _pb2.py files with nanopb options
+# embedded in the serialized descriptors.
+# ===========================================================================
+
+from meshtastic.protobuf import (  # noqa: E402  (after local helpers)
+    atak_pb2,
+    config_pb2,
+    mesh_pb2,
+    nanopb_pb2,
+    telemetry_pb2,
+)
+
+
+def _field_opts(descriptor, *path):
+    """Walk a descriptor by field/nested-type path and return its nanopb opts.
+
+    Elements of *path that are message names are looked up in nested_types_by_name;
+    the final element is looked up in fields_by_name.
+    """
+    desc = descriptor
+    for step in path[:-1]:
+        desc = desc.nested_types_by_name[step]
+    field = desc.fields_by_name[path[-1]]
+    return field.GetOptions().Extensions[nanopb_pb2.nanopb]
+
+
+@pytest.mark.unit
+def test_descriptor_user_long_name():
+    opts = _field_opts(mesh_pb2.DESCRIPTOR.message_types_by_name["User"], "long_name")
+    assert opts.max_size == 40
+
+
+@pytest.mark.unit
+def test_descriptor_user_short_name():
+    opts = _field_opts(mesh_pb2.DESCRIPTOR.message_types_by_name["User"], "short_name")
+    assert opts.max_size == 5
+
+
+@pytest.mark.unit
+def test_descriptor_wildcard_macaddr():
+    """Wildcard option from mesh.options applied to User.macaddr."""
+    opts = _field_opts(mesh_pb2.DESCRIPTOR.message_types_by_name["User"], "macaddr")
+    assert opts.max_size == 6
+    assert opts.fixed_length is True
+
+
+@pytest.mark.unit
+def test_descriptor_meshpacket_hop_limit():
+    opts = _field_opts(mesh_pb2.DESCRIPTOR.message_types_by_name["MeshPacket"], "hop_limit")
+    assert opts.int_size == nanopb_pb2.IS_8
+
+
+@pytest.mark.unit
+def test_descriptor_routediscovery_snr_towards():
+    opts = _field_opts(
+        mesh_pb2.DESCRIPTOR.message_types_by_name["RouteDiscovery"], "snr_towards"
+    )
+    assert opts.max_count == 8
+    assert opts.int_size == nanopb_pb2.IS_8
+
+
+@pytest.mark.unit
+def test_descriptor_data_payload():
+    opts = _field_opts(mesh_pb2.DESCRIPTOR.message_types_by_name["Data"], "payload")
+    assert opts.max_size == 233
+
+
+@pytest.mark.unit
+def test_descriptor_nested_deviceconfig_tzdef():
+    """Config.DeviceConfig.tzdef — option on a field inside a nested message."""
+    config = config_pb2.DESCRIPTOR.message_types_by_name["Config"]
+    opts = _field_opts(config, "DeviceConfig", "tzdef")
+    assert opts.max_size == 65
+
+
+@pytest.mark.unit
+def test_descriptor_nested_securityconfig_admin_key():
+    """Config.SecurityConfig.admin_key — two options merged from two .options lines."""
+    config = config_pb2.DESCRIPTOR.message_types_by_name["Config"]
+    opts = _field_opts(config, "SecurityConfig", "admin_key")
+    assert opts.max_size == 32
+    assert opts.max_count == 3
+
+
+@pytest.mark.unit
+def test_descriptor_multilevel_nested_route_link_uid():
+    """Route.Link.uid — three-level nested pattern from atak.options."""
+    route = atak_pb2.DESCRIPTOR.message_types_by_name["Route"]
+    opts = _field_opts(route, "Link", "uid")
+    assert opts.max_size == 48
+
+
+@pytest.mark.unit
+def test_descriptor_telemetry_environment_one_wire_temperature():
+    env = telemetry_pb2.DESCRIPTOR.message_types_by_name["EnvironmentMetrics"]
+    opts = _field_opts(env, "one_wire_temperature")
+    assert opts.max_count == 8