diff --git a/dev-notes/resource-diagnostics-command.md b/dev-notes/resource-diagnostics-command.md
new file mode 100644
index 000000000..a778ca838
--- /dev/null
+++ b/dev-notes/resource-diagnostics-command.md
@@ -0,0 +1,29 @@
+# Resource diagnostics command
+
+Issue #459 exposes Tau's existing non-fatal resource diagnostics through a
+built-in `/diagnostics` command. The command reads
+`CodingSession.resource_diagnostics` when invoked, so its output reflects both
+initial resource discovery and the latest `/reload`.
+
+The shared command handler renders an explicit empty state or every diagnostic's
+severity, resource kind, optional name, optional source path, and message. Print
+mode writes the same text directly. The Textual frontend uses its existing
+scrollable command-output modal, which already renders command messages with
+markup disabled, so paths and error text containing Rich/Textual markup
+characters remain literal.
+
+`diagnostics.md` is a reserved prompt-template name. Ignoring a colliding
+template prevents prompt expansion from shadowing the built-in command and
+records a diagnostic explaining the collision.
+
+Validate manually:
+
+1. Add an invalid skill, prompt, context file, or extension.
+2. Start Tau and run `/diagnostics`.
+3. Confirm the modal shows the resource kind, source path, severity, and
+   remediation message.
+4. Fix the resource, run `/reload`, and run `/diagnostics` again.
+5. Confirm the modal shows `No resource diagnostics.`
+
+Automated coverage lives in `tests/test_commands.py`, `tests/test_cli.py`,
+`tests/test_prompt_templates.py`, and `tests/test_tui_app.py`.
diff --git a/src/tau_coding/commands.py b/src/tau_coding/commands.py
index e297c837b..4f8622fd9 100644
--- a/src/tau_coding/commands.py
+++ b/src/tau_coding/commands.py
@@ -250,6 +250,15 @@ def create_default_command_registry() -> CommandRegistry:
             search_terms=("info",),
         )
     )
+    registry.register(
+        SlashCommand(
+            name="diagnostics",
+            usage="/diagnostics",
+            description="Show resource loading diagnostics.",
+            handler=_diagnostics_command,
+            search_terms=("errors", "warnings", "loading"),
+        )
+    )
     registry.register(
         SlashCommand(
             name="system",
@@ -486,19 +495,13 @@ def _skills_command(context: CommandContext) -> CommandResult:
     return CommandResult(handled=True, skills_picker_requested=True)
 
 
-def _resources_command(context: CommandContext) -> CommandResult:
-    session = context.session
-    lines = [
-        f"Skills: {len(session.skills)}",
-        f"Prompt templates: {len(session.prompt_templates)}",
-        f"Context files: {len(session.context_files)}",
-    ]
-    if session.resource_diagnostics:
-        lines.append("")
-        lines.extend(_format_diagnostics(session.resource_diagnostics))
-    else:
-        lines.append("Resource diagnostics: none")
-    return CommandResult(handled=True, message="\n".join(lines))
+def _diagnostics_command(context: CommandContext) -> CommandResult:
+    if context.args:
+        return CommandResult(handled=True, message="Usage: /diagnostics")
+    return CommandResult(
+        handled=True,
+        message=format_resource_diagnostics(context.session.resource_diagnostics),
+    )
 
 
 def _reload_command(context: CommandContext) -> CommandResult:
@@ -780,6 +783,28 @@ def _format_diagnostics(
     return lines
 
 
+def format_resource_diagnostics(diagnostics: Sequence[ResourceDiagnostic]) -> str:
+    """Render complete resource diagnostics for command output."""
+    if not diagnostics:
+        return "No resource diagnostics."
+
+    lines = [f"Resource diagnostics ({len(diagnostics)}):"]
+    for index, diagnostic in enumerate(diagnostics, start=1):
+        lines.extend(
+            [
+                "",
+                f"{index}. {diagnostic.severity.upper()}",
+                f"Kind: {diagnostic.kind}",
+            ]
+        )
+        if diagnostic.name is not None:
+            lines.append(f"Name: {diagnostic.name}")
+        if diagnostic.path is not None:
+            lines.append(f"Path: {diagnostic.path}")
+        lines.append(f"Message: {diagnostic.message}")
+    return "\n".join(lines)
+
+
 def _refresh_provider_settings(session: CommandSession) -> CommandResult | None:
     try:
         session.reload_provider_settings()
diff --git a/src/tau_coding/prompt_templates.py b/src/tau_coding/prompt_templates.py
index 24fdb05d0..5af470006 100644
--- a/src/tau_coding/prompt_templates.py
+++ b/src/tau_coding/prompt_templates.py
@@ -17,7 +17,7 @@
 
 _TEMPLATE_VARIABLE_RE = re.compile(r"{{\s*([a-zA-Z_][a-zA-Z0-9_]*)\s*}}")
 _ARGUMENT_TEMPLATE_VARIABLES = {"arguments", "args"}
-_RESERVED_TEMPLATE_NAMES = frozenset({"prompts", "skills", "tools"})
+_RESERVED_TEMPLATE_NAMES = frozenset({"diagnostics", "prompts", "skills", "tools"})
 
 
 @dataclass(frozen=True, slots=True)
diff --git a/tests/test_cli.py b/tests/test_cli.py
index 83c69edd8..caf9e6eb6 100644
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -450,6 +450,46 @@ async def test_run_print_mode_system_command_prints_prompt_without_provider_call
     assert await storage.read_all() == []
 
 
+@pytest.mark.anyio
+async def test_run_print_mode_diagnostics_command_prints_resource_details(
+    capsys: pytest.CaptureFixture[str], tmp_path: Path
+) -> None:
+    resource_root = tmp_path / "resources"
+    prompts_dir = resource_root / "prompts"
+    prompts_dir.mkdir(parents=True)
+    diagnostic_path = prompts_dir / "diagnostics.md"
+    diagnostic_path.write_text("Shadow the built-in command.", encoding="utf-8")
+    provider = FakeProvider([])
+
+    ok = await run_print_mode(
+        prompt="/diagnostics",
+        model="fake",
+        cwd=tmp_path,
+        provider=provider,
+        resource_paths=TauResourcePaths(root=resource_root, agents_root=None),
+    )
+
+    captured = capsys.readouterr()
+    assert ok is True
+    assert captured.out == "\n".join(
+        [
+            "Resource diagnostics (1):",
+            "",
+            "1. WARNING",
+            "Kind: prompt",
+            "Name: diagnostics",
+            f"Path: {diagnostic_path}",
+            (
+                "Message: prompt template name is reserved by the built-in "
+                "/diagnostics command; template ignored"
+            ),
+            "",
+        ]
+    )
+    assert captured.err == ""
+    assert provider.calls == []
+
+
 @pytest.mark.anyio
 async def test_run_print_mode_fails_on_non_recoverable_error(
     capsys: pytest.CaptureFixture[str], tmp_path: Path
diff --git a/tests/test_commands.py b/tests/test_commands.py
index 62eb64b58..9bf528e34 100644
--- a/tests/test_commands.py
+++ b/tests/test_commands.py
@@ -3,6 +3,7 @@
 from tau_coding.commands import CommandRegistry, SlashCommand, create_default_command_registry
 from tau_coding.paths import TauPaths
 from tau_coding.reload import CodingReloadSummary, ReloadCategorySummary
+from tau_coding.resources import ResourceDiagnostic
 from tau_coding.session import ModelChoice
 from tau_coding.session_manager import SessionManager
 from tau_coding.skills import Skill
@@ -114,11 +115,12 @@ def test_registry_ignores_unregistered_slash_prompts(tmp_path: Path) -> None:
         assert result.message is None
 
 
-def test_registered_commands_are_pi_aligned(tmp_path: Path) -> None:
+def test_registered_commands_are_expected(tmp_path: Path) -> None:
     commands = create_default_command_registry().list_commands()
 
     assert [command.name for command in commands] == [
         "compact",
+        "diagnostics",
         "export",
         "hotkeys",
         "login",
@@ -270,6 +272,56 @@ def test_session_command_explains_unavailable_thinking_controls(tmp_path: Path)
     assert "Thinking mode: medium" not in result.message
 
 
+def test_diagnostics_command_shows_empty_state(tmp_path: Path) -> None:
+    result = create_default_command_registry().execute(FakeSession(tmp_path), "/diagnostics")
+
+    assert result.handled is True
+    assert result.message == "No resource diagnostics."
+    assert (
+        create_default_command_registry()
+        .execute(FakeSession(tmp_path), "/diagnostics extra")
+        .message
+        == "Usage: /diagnostics"
+    )
+
+
+def test_diagnostics_command_shows_complete_literal_details(tmp_path: Path) -> None:
+    session = FakeSession(tmp_path)
+    prompt_path = tmp_path / "[project]" / "prompts" / "review.md"
+    session.resource_diagnostics = (
+        ResourceDiagnostic(
+            kind="prompt",
+            name="review",
+            path=prompt_path,
+            severity="warning",
+            message="template [ignored] because it is invalid",
+        ),
+        ResourceDiagnostic(
+            kind="extension",
+            severity="error",
+            message="setup failed",
+        ),
+    )
+
+    result = create_default_command_registry().execute(session, "/diagnostics")
+
+    assert result.message == "\n".join(
+        [
+            "Resource diagnostics (2):",
+            "",
+            "1. WARNING",
+            "Kind: prompt",
+            "Name: review",
+            f"Path: {prompt_path}",
+            "Message: template [ignored] because it is invalid",
+            "",
+            "2. ERROR",
+            "Kind: extension",
+            "Message: setup failed",
+        ]
+    )
+
+
 def test_hotkeys_command_lists_common_tui_shortcuts(tmp_path: Path) -> None:
     result = create_default_command_registry().execute(FakeSession(tmp_path), "/hotkeys")
 
diff --git a/tests/test_prompt_templates.py b/tests/test_prompt_templates.py
index 8886e2169..6440d8a4f 100644
--- a/tests/test_prompt_templates.py
+++ b/tests/test_prompt_templates.py
@@ -90,7 +90,7 @@ def test_load_prompt_templates_with_diagnostics_reports_overrides(tmp_path: Path
     assert "overrides lower-precedence resource" in diagnostics[0].message
 
 
-@pytest.mark.parametrize("reserved_name", ["prompts", "skills", "tools"])
+@pytest.mark.parametrize("reserved_name", ["diagnostics", "prompts", "skills", "tools"])
 def test_reserved_picker_template_is_ignored_with_diagnostic(
     tmp_path: Path, reserved_name: str
 ) -> None:
diff --git a/tests/test_tui_app.py b/tests/test_tui_app.py
index 31e439675..7b9c5abf9 100644
--- a/tests/test_tui_app.py
+++ b/tests/test_tui_app.py
@@ -42,7 +42,7 @@
 from tau_agent.messages import assistant_content
 from tau_agent.provider_events import TextDeltaEvent, ThinkingDeltaEvent
 from tau_coding.catalog_loader import user_catalog_path
-from tau_coding.commands import CommandResult
+from tau_coding.commands import CommandResult, format_resource_diagnostics
 from tau_coding.credentials import FileCredentialStore, OAuthCredential
 from tau_coding.events import AgentSettledEvent, CodingSessionEvent, QueueUpdateEvent
 from tau_coding.paths import TauPaths
@@ -54,6 +54,7 @@
     ScopedModelConfig,
     save_provider_settings,
 )
+from tau_coding.resources import ResourceDiagnostic
 from tau_coding.session import (
     ModelChoice,
     SessionTreeBranchResult,
@@ -241,6 +242,11 @@ def handle_command(self, text: str) -> CommandResult:
             )
         if text == "/system":
             return CommandResult(handled=True, message=self.system_prompt)
+        if text == "/diagnostics":
+            return CommandResult(
+                handled=True,
+                message=format_resource_diagnostics(self.resource_diagnostics),
+            )
         if text == "/skills":
             return CommandResult(handled=True, skills_picker_requested=True)
         if text == "/new":
@@ -5226,6 +5232,7 @@ async def test_tui_app_opens_command_palette_from_keybinding() -> None:
         assert prompt.value == "/"
         assert app._completion_state.items
         assert any(item.display == "/session" for item in app._completion_state.items)
+        assert any(item.display == "/diagnostics" for item in app._completion_state.items)
         assert app.query_one("#autocomplete").display is True
 
 
@@ -5321,6 +5328,80 @@ async def test_tui_app_help_uses_modal_instead_of_transcript() -> None:
         assert app.screen.focused is scroll
 
 
+@pytest.mark.anyio
+async def test_tui_app_diagnostics_modal_shows_empty_state_and_dismisses() -> None:
+    app = TauTuiApp(FakeSession())
+
+    async with app.run_test() as pilot:
+        prompt = app.query_one("#prompt", PromptInput)
+        prompt.value = "/diagnostics"
+        await pilot.press("enter")
+        await pilot.pause()
+
+        assert isinstance(app.screen, CommandOutputScreen)
+        assert app.screen.title_text == "/diagnostics"
+        assert app.screen.message == "No resource diagnostics."
+        assert app.state.items == []
+
+        await pilot.press("escape")
+        await pilot.pause()
+
+        assert not isinstance(app.screen, CommandOutputScreen)
+        assert prompt.value == ""
+
+
+@pytest.mark.anyio
+async def test_tui_app_diagnostics_modal_reflects_post_reload_details() -> None:
+    diagnostic_path = Path("/tmp/[project]/prompts/review.md")
+    diagnostic = ResourceDiagnostic(
+        kind="prompt",
+        name="review",
+        path=diagnostic_path,
+        severity="warning",
+        message="template [ignored]; rename it and reload",
+    )
+
+    class ReloadingDiagnosticSession(FakeSession):
+        def handle_command(self, text: str) -> CommandResult:
+            if text == "/reload":
+                self.resource_diagnostics = (diagnostic,)
+                return CommandResult(handled=True, message="Reloaded resources.")
+            return super().handle_command(text)
+
+    app = TauTuiApp(ReloadingDiagnosticSession())
+
+    async with app.run_test() as pilot:
+        prompt = app.query_one("#prompt", PromptInput)
+        prompt.value = "/diagnostics"
+        await pilot.press("enter")
+        await pilot.pause()
+        assert isinstance(app.screen, CommandOutputScreen)
+        assert app.screen.message == "No resource diagnostics."
+
+        await pilot.press("escape")
+        prompt.value = "/reload"
+        await pilot.press("enter")
+        await pilot.pause()
+
+        prompt.value = "/diagnostics"
+        await pilot.press("enter")
+        await pilot.pause()
+
+        assert isinstance(app.screen, CommandOutputScreen)
+        body = app.screen.query_one("#command-output-body", Static)
+        assert str(body.render()) == "\n".join(
+            [
+                "Resource diagnostics (1):",
+                "",
+                "1. WARNING",
+                "Kind: prompt",
+                "Name: review",
+                f"Path: {diagnostic_path}",
+                "Message: template [ignored]; rename it and reload",
+            ]
+        )
+
+
 @pytest.mark.anyio
 async def test_tui_app_tools_reference_opens_filters_and_cancels() -> None:
     app = TauTuiApp(FakeSession())
diff --git a/website/content/guides/tui.md b/website/content/guides/tui.md
index 19d9ebc87..2e6944ddc 100644
--- a/website/content/guides/tui.md
+++ b/website/content/guides/tui.md
@@ -55,6 +55,7 @@ In-session commands start with `/`. Open the **command palette** with **Ctrl+K**
 to search and run them. Common ones:
 
 - `/session` — show model, tools, skills, and context usage for the session. Text selected in this modal is copied to the clipboard automatically.
+- `/diagnostics` — inspect resource loading warnings and errors, including the resource kind, name, source path, severity, and remediation message
 - `/model` — pick the active model
 - `/tools` — search active tools by origin and open their full descriptions
 - `/compact` — summarize and shrink the context
diff --git a/website/content/reference/slash-commands.md b/website/content/reference/slash-commands.md
index dc24eccb3..767ec699e 100644
--- a/website/content/reference/slash-commands.md
+++ b/website/content/reference/slash-commands.md
@@ -11,6 +11,7 @@ command palette with **Ctrl+K**.
 | `/quit` | Exit the session |
 | `/new` | Start a new session |
 | `/session` | Show session info and stats (model, cwd, tools, skills, context) |
+| `/diagnostics` | Inspect resource loading warnings and errors, including their source paths |
 | `/system` | Show the active system prompt without adding it to context or session history |
 | `/compact [instructions]` | Summarize and compact the active context |
 | `/export [--format html\|jsonl] [dest]` | Export the current session |