diff --git a/packages/mcp/src/pipefy_mcp/tools/ai_agent_tools.py b/packages/mcp/src/pipefy_mcp/tools/ai_agent_tools.py
index 5cb183b6..2adf1b95 100644
--- a/packages/mcp/src/pipefy_mcp/tools/ai_agent_tools.py
+++ b/packages/mcp/src/pipefy_mcp/tools/ai_agent_tools.py
@@ -41,6 +41,7 @@
     enrich_permission_denied_error,
     extract_error_strings,
 )
+from pipefy_mcp.tools.remote_profile import REMOTE
 from pipefy_mcp.tools.tool_context import get_pipefy_client
 
 VALIDATE_FETCH_TIMEOUT_SECONDS = 30
@@ -481,6 +482,7 @@ async def toggle_ai_agent_status(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_ai_agent(ctx: Context, uuid: str) -> dict:
             """Get an AI Agent by UUID with full behavior configuration.
@@ -513,6 +515,7 @@ async def get_ai_agent(ctx: Context, uuid: str) -> dict:
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_ai_agents(ctx: Context, repo_uuid: str) -> dict:
             """List all AI Agents for a pipe. Use before creating an agent to avoid duplicates.
@@ -575,6 +578,7 @@ async def delete_ai_agent(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True, destructiveHint=False),
+            meta=REMOTE,
         )
         async def validate_ai_agent_behaviors(
             ctx: Context,
diff --git a/packages/mcp/src/pipefy_mcp/tools/ai_automation_tools.py b/packages/mcp/src/pipefy_mcp/tools/ai_automation_tools.py
index 7c0a77f3..df795ab5 100644
--- a/packages/mcp/src/pipefy_mcp/tools/ai_automation_tools.py
+++ b/packages/mcp/src/pipefy_mcp/tools/ai_automation_tools.py
@@ -29,6 +29,7 @@
     handle_automation_tool_graphql_error,
 )
 from pipefy_mcp.tools.destructive_tool_guard import check_destructive_confirmation
+from pipefy_mcp.tools.remote_profile import REMOTE
 from pipefy_mcp.tools.tool_context import get_pipefy_client
 from pipefy_mcp.tools.validation_helpers import (
     validate_optional_tool_id,
@@ -45,6 +46,7 @@ def register(mcp: FastMCP) -> None:
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True, destructiveHint=False),
+            meta=REMOTE,
         )
         async def validate_ai_automation_prompt(
             ctx: Context,
@@ -125,6 +127,7 @@ async def validate_ai_automation_prompt(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True, destructiveHint=False),
+            meta=REMOTE,
         )
         async def get_ai_automation(
             ctx: Context,
@@ -173,6 +176,7 @@ async def get_ai_automation(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True, destructiveHint=False),
+            meta=REMOTE,
         )
         async def get_ai_automations(
             ctx: Context,
diff --git a/packages/mcp/src/pipefy_mcp/tools/llm_provider_tools.py b/packages/mcp/src/pipefy_mcp/tools/llm_provider_tools.py
index 695665f2..12debe03 100644
--- a/packages/mcp/src/pipefy_mcp/tools/llm_provider_tools.py
+++ b/packages/mcp/src/pipefy_mcp/tools/llm_provider_tools.py
@@ -17,6 +17,7 @@
     build_pagination_info,
     validate_page_size,
 )
+from pipefy_mcp.tools.remote_profile import REMOTE
 from pipefy_mcp.tools.tool_context import get_pipefy_client
 from pipefy_mcp.tools.validation_helpers import validate_tool_id
 
@@ -132,7 +133,7 @@ async def get_llm_providers(
                 page_size=nfirst,
             )
 
-        @mcp.tool(annotations=ToolAnnotations(readOnlyHint=True))
+        @mcp.tool(annotations=ToolAnnotations(readOnlyHint=True), meta=REMOTE)
         async def get_available_ai_models(
             ctx: Context,
             provider_name: str,
diff --git a/packages/mcp/src/pipefy_mcp/tools/observability_tools.py b/packages/mcp/src/pipefy_mcp/tools/observability_tools.py
index 9457de8d..882d9d87 100644
--- a/packages/mcp/src/pipefy_mcp/tools/observability_tools.py
+++ b/packages/mcp/src/pipefy_mcp/tools/observability_tools.py
@@ -22,6 +22,7 @@
     build_observability_read_success_payload,
     handle_observability_tool_graphql_error,
 )
+from pipefy_mcp.tools.remote_profile import REMOTE
 from pipefy_mcp.tools.tool_context import get_pipefy_client
 from pipefy_mcp.tools.validation_helpers import (
     validate_optional_tool_id,
@@ -72,6 +73,7 @@ class ObservabilityTools:
     def register(mcp: FastMCP) -> None:
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_ai_agent_logs(
             repo_uuid: str,
@@ -123,6 +125,7 @@ async def get_ai_agent_logs(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_ai_agent_log_details(
             log_uuid: str,
@@ -261,6 +264,7 @@ async def get_automation_logs_by_repo(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_agents_usage(
             organization_uuid: PipefyId,
@@ -369,6 +373,7 @@ async def get_automations_usage(
 
         @mcp.tool(
             annotations=ToolAnnotations(readOnlyHint=True),
+            meta=REMOTE,
         )
         async def get_ai_credit_usage(
             organization_uuid: PipefyId,
diff --git a/packages/mcp/tests/tools/test_remote_profile.py b/packages/mcp/tests/tools/test_remote_profile.py
index 3cbb5a4c..ee593e6a 100644
--- a/packages/mcp/tests/tools/test_remote_profile.py
+++ b/packages/mcp/tests/tools/test_remote_profile.py
@@ -39,6 +39,20 @@
         "introspect_query",
         "introspect_mutation",
         "introspect_type",
+        # AI agent & automation reads (#437): read/validate tools that reach the
+        # API with the request-scoped bearer and are governed by API permissions;
+        # no filesystem or per-user process-global settings reads.
+        "get_ai_agent",
+        "get_ai_agents",
+        "get_ai_agent_logs",
+        "get_ai_agent_log_details",
+        "get_agents_usage",
+        "validate_ai_agent_behaviors",
+        "get_ai_automation",
+        "get_ai_automations",
+        "validate_ai_automation_prompt",
+        "get_ai_credit_usage",
+        "get_available_ai_models",
     }
 )