claude-mpm 4.25.10__py3-none-any.whl → 5.1.8__py3-none-any.whl

claude_mpm/services/diagnostics/checks/agent_sources_check.py

@@ -0,0 +1,577 @@
+"""
+Check agent sources configuration and health.
+
+WHY: Verify that the single-tier Git-based agent system is properly configured,
+sources are accessible, and agents are discoverable from configured repositories.
+"""
+
+import logging
+from pathlib import Path
+from typing import Dict, List, Optional
+
+from ....core.enums import OperationResult, ValidationSeverity
+from ..models import DiagnosticResult
+from .base_check import BaseDiagnosticCheck
+
+logger = logging.getLogger(__name__)
+
+
+class AgentSourcesCheck(BaseDiagnosticCheck):
+    """Check agent sources configuration and health.
+
+    Verifies:
+    - Configuration file exists and is valid YAML
+    - At least one source is configured
+    - Enabled sources are accessible (HTTP 200)
+    - Cache directory is healthy and writable
+    - Priority conflicts are identified
+    - Agents can be discovered from sources
+
+    WHY: The single-tier system relies on Git sources being properly configured
+    and accessible. This check ensures the agent deployment pipeline is healthy.
+    """
+
+    @property
+    def name(self) -> str:
+        return "agent_sources_check"
+
+    @property
+    def category(self) -> str:
+        return "Agent Sources"
+
+    def should_run(self) -> bool:
+        """Check should run if single-tier system is in use.
+
+        Returns:
+            True if agent sources configuration exists or could exist
+        """
+        # Always run - we want to check if configuration should be created
+        return True
+
+    def run(self) -> DiagnosticResult:
+        """Run agent sources diagnostics."""
+        try:
+            sub_results = []
+            details = {}
+
+            # Check 1: Configuration file exists
+            config_result = self._check_config_file()
+            sub_results.append(config_result)
+            details["config_exists"] = config_result.status == OperationResult.SUCCESS
+
+            # If no config file, return early with guidance
+            if config_result.status == ValidationSeverity.ERROR:
+                return DiagnosticResult(
+                    category=self.category,
+                    status=ValidationSeverity.ERROR,
+                    message="Agent sources not configured",
+                    details=details,
+                    fix_command="claude-mpm agent-source add https://github.com/bobmatnyc/claude-mpm-agents",
+                    fix_description="Add default system repository",
+                    sub_results=sub_results if self.verbose else [],
+                )
+
+            # Check 2: Configuration is valid YAML
+            config_valid_result = self._check_config_valid()
+            sub_results.append(config_valid_result)
+            if config_valid_result.status == ValidationSeverity.ERROR:
+                return DiagnosticResult(
+                    category=self.category,
+                    status=ValidationSeverity.ERROR,
+                    message="Invalid configuration file",
+                    details=details,
+                    fix_command="claude-mpm config validate",
+                    fix_description="Validate and fix configuration",
+                    sub_results=sub_results if self.verbose else [],
+                )
+
+            # Load configuration for remaining checks
+            from ....config.agent_sources import AgentSourceConfiguration
+
+            config = AgentSourceConfiguration.load()
+
+            # Check 3: At least one source configured
+            sources_result = self._check_sources_configured(config)
+            sub_results.append(sources_result)
+            details["total_sources"] = len(config.repositories) + (
+                0 if config.disable_system_repo else 1
+            )
+            details["enabled_sources"] = len(config.get_enabled_repositories())
+
+            # Check 4: System repository accessible (if enabled)
+            system_repo_result = self._check_system_repo_accessible(config)
+            if system_repo_result:
+                sub_results.append(system_repo_result)
+
+            # Check 5: Enabled sources reachable
+            reachability_result = self._check_sources_reachable(config)
+            sub_results.append(reachability_result)
+            details["reachable_sources"] = reachability_result.details.get(
+                "reachable_count", 0
+            )
+            details["unreachable_sources"] = reachability_result.details.get(
+                "unreachable_count", 0
+            )
+
+            # Check 6: Cache directory healthy
+            cache_result = self._check_cache_directory()
+            sub_results.append(cache_result)
+            details["cache_healthy"] = cache_result.status == OperationResult.SUCCESS
+
+            # Check 7: Priority conflicts
+            priority_result = self._check_priority_conflicts(config)
+            if priority_result:
+                sub_results.append(priority_result)
+                details["priority_conflicts"] = priority_result.details.get(
+                    "conflict_count", 0
+                )
+
+            # Check 8: Agents discovered
+            discovery_result = self._check_agents_discovered(config)
+            sub_results.append(discovery_result)
+            details["agents_discovered"] = discovery_result.details.get(
+                "total_agents", 0
+            )
+
+            # Determine overall status
+            error_results = [
+                r for r in sub_results if r.status == ValidationSeverity.ERROR
+            ]
+            warning_results = [
+                r for r in sub_results if r.status == ValidationSeverity.WARNING
+            ]
+
+            if error_results:
+                status = ValidationSeverity.ERROR
+                message = f"Agent sources have {len(error_results)} critical issue(s)"
+                fix_command = None
+                fix_description = None
+            elif warning_results:
+                status = ValidationSeverity.WARNING
+                message = f"Agent sources have {len(warning_results)} minor issue(s)"
+                fix_command = "claude-mpm agent-source update"
+                fix_description = "Update all sources to refresh cache"
+            else:
+                status = OperationResult.SUCCESS
+                enabled_count = details["enabled_sources"]
+                agents_count = details["agents_discovered"]
+                message = f"All checks passed ({enabled_count} source(s), {agents_count} agent(s))"
+                fix_command = None
+                fix_description = None
+
+            return DiagnosticResult(
+                category=self.category,
+                status=status,
+                message=message,
+                details=details,
+                fix_command=fix_command,
+                fix_description=fix_description,
+                sub_results=sub_results if self.verbose else [],
+            )
+
+        except Exception as e:
+            logger.exception("Agent sources check failed")
+            return DiagnosticResult(
+                category=self.category,
+                status=ValidationSeverity.ERROR,
+                message=f"Agent sources check failed: {e!s}",
+                details={"error": str(e)},
+            )
+
+    def _check_config_file(self) -> DiagnosticResult:
+        """Check if configuration file exists."""
+        config_path = Path.home() / ".claude-mpm" / "config" / "agent_sources.yaml"
+
+        if not config_path.exists():
+            return DiagnosticResult(
+                category="Configuration File",
+                status=ValidationSeverity.ERROR,
+                message=f"Configuration file not found: {config_path}",
+                details={"path": str(config_path)},
+                fix_command="claude-mpm agent-source add https://github.com/bobmatnyc/claude-mpm-agents",
+                fix_description="Create configuration and add default repository",
+            )
+
+        return DiagnosticResult(
+            category="Configuration File",
+            status=OperationResult.SUCCESS,
+            message=f"Found at {config_path}",
+            details={"path": str(config_path)},
+        )
+
+    def _check_config_valid(self) -> DiagnosticResult:
+        """Check if configuration is valid YAML."""
+        config_path = Path.home() / ".claude-mpm" / "config" / "agent_sources.yaml"
+
+        try:
+            import yaml
+
+            with open(config_path) as f:
+                data = yaml.safe_load(f)
+
+            if not data:
+                return DiagnosticResult(
+                    category="Configuration Validity",
+                    status=ValidationSeverity.WARNING,
+                    message="Configuration file is empty",
+                    details={"path": str(config_path)},
+                )
+
+            # Basic structure validation
+            if not isinstance(data, dict):
+                return DiagnosticResult(
+                    category="Configuration Validity",
+                    status=ValidationSeverity.ERROR,
+                    message="Configuration must be YAML dictionary",
+                    details={"path": str(config_path), "type": str(type(data))},
+                )
+
+            return DiagnosticResult(
+                category="Configuration Validity",
+                status=OperationResult.SUCCESS,
+                message="Configuration is valid YAML",
+                details={"path": str(config_path)},
+            )
+
+        except yaml.YAMLError as e:
+            return DiagnosticResult(
+                category="Configuration Validity",
+                status=ValidationSeverity.ERROR,
+                message=f"Invalid YAML: {e!s}",
+                details={"path": str(config_path), "error": str(e)},
+                fix_command=f"cat {config_path}",
+                fix_description="Check YAML syntax errors",
+            )
+        except Exception as e:
+            return DiagnosticResult(
+                category="Configuration Validity",
+                status=ValidationSeverity.ERROR,
+                message=f"Failed to read configuration: {e!s}",
+                details={"path": str(config_path), "error": str(e)},
+            )
+
+    def _check_sources_configured(self, config) -> DiagnosticResult:
+        """Check if at least one source is configured.
+
+        Args:
+            config: AgentSourceConfiguration instance
+        """
+        total_sources = len(config.repositories)
+        if not config.disable_system_repo:
+            total_sources += 1  # Count system repo
+
+        if total_sources == 0:
+            return DiagnosticResult(
+                category="Sources Configured",
+                status=ValidationSeverity.WARNING,
+                message="No agent sources configured",
+                details={"total": 0, "enabled": 0},
+                fix_command="claude-mpm agent-source add https://github.com/bobmatnyc/claude-mpm-agents",
+                fix_description="Add default agent repository",
+            )
+
+        enabled_sources = len(config.get_enabled_repositories())
+
+        if enabled_sources == 0:
+            return DiagnosticResult(
+                category="Sources Configured",
+                status=ValidationSeverity.WARNING,
+                message=f"{total_sources} source(s) configured but all disabled",
+                details={"total": total_sources, "enabled": 0},
+                fix_command="claude-mpm source enable <source-id>",
+                fix_description="Enable at least one source",
+            )
+
+        return DiagnosticResult(
+            category="Sources Configured",
+            status=OperationResult.SUCCESS,
+            message=f"{enabled_sources} enabled source(s) ({total_sources} total)",
+            details={"total": total_sources, "enabled": enabled_sources},
+        )
+
+    def _check_system_repo_accessible(self, config) -> Optional[DiagnosticResult]:
+        """Check if system repository is accessible.
+
+        Args:
+            config: AgentSourceConfiguration instance
+        """
+        system_repo = config.get_system_repo()
+
+        if not system_repo:
+            # System repo disabled, skip check
+            return None
+
+        return self._check_repo_accessible(
+            system_repo.url, "System Repository", is_system=True
+        )
+
+    def _check_sources_reachable(self, config) -> DiagnosticResult:
+        """Check if enabled sources are reachable.
+
+        Args:
+            config: AgentSourceConfiguration instance
+        """
+        repos = config.get_enabled_repositories()
+
+        if not repos:
+            return DiagnosticResult(
+                category="Source Reachability",
+                status=OperationResult.SKIPPED,
+                message="No enabled sources to check",
+                details={"reachable_count": 0, "unreachable_count": 0},
+            )
+
+        reachable = []
+        unreachable = []
+        details: Dict[str, str] = {}
+
+        for repo in repos:
+            result = self._check_repo_accessible(repo.url, repo.identifier)
+            if result.status == OperationResult.SUCCESS:
+                reachable.append(repo.identifier)
+                details[repo.identifier] = "reachable"
+            else:
+                unreachable.append(repo.identifier)
+                details[repo.identifier] = result.message
+
+        if len(unreachable) == len(repos):
+            # All sources unreachable
+            return DiagnosticResult(
+                category="Source Reachability",
+                status=ValidationSeverity.ERROR,
+                message=f"All {len(repos)} source(s) unreachable",
+                details={
+                    "reachable_count": 0,
+                    "unreachable_count": len(unreachable),
+                    "sources": details,
+                },
+                fix_command="claude-mpm agent-source list",
+                fix_description="Check source URLs and network connectivity",
+            )
+        if unreachable:
+            # Some sources unreachable
+            return DiagnosticResult(
+                category="Source Reachability",
+                status=ValidationSeverity.WARNING,
+                message=f"{len(unreachable)}/{len(repos)} source(s) unreachable",
+                details={
+                    "reachable_count": len(reachable),
+                    "unreachable_count": len(unreachable),
+                    "sources": details,
+                    "unreachable_sources": unreachable,
+                },
+            )
+        # All sources reachable
+        return DiagnosticResult(
+            category="Source Reachability",
+            status=OperationResult.SUCCESS,
+            message=f"All {len(repos)} source(s) reachable",
+            details={
+                "reachable_count": len(reachable),
+                "unreachable_count": 0,
+                "sources": details,
+            },
+        )
+
+    def _check_repo_accessible(
+        self, url: str, identifier: str, is_system: bool = False
+    ) -> DiagnosticResult:
+        """Check if a repository is accessible via HTTP."""
+        try:
+            import urllib.request
+            from urllib.error import HTTPError, URLError
+
+            # Convert git URL to HTTPS if needed
+            if url.startswith("git@github.com:"):
+                url = url.replace("git@github.com:", "https://github.com/")
+            if url.endswith(".git"):
+                url = url[:-4]
+
+            # Try to fetch repository (use HEAD request to be lightweight)
+            req = urllib.request.Request(url, method="HEAD")
+            req.add_header("User-Agent", "claude-mpm-doctor/1.0")
+
+            with urllib.request.urlopen(req, timeout=5) as response:
+                status = response.status
+
+            if status == 200:
+                return DiagnosticResult(
+                    category=identifier,
+                    status=OperationResult.SUCCESS,
+                    message="Accessible",
+                    details={"url": url, "status": status},
+                )
+            return DiagnosticResult(
+                category=identifier,
+                status=ValidationSeverity.WARNING,
+                message=f"HTTP {status}",
+                details={"url": url, "status": status},
+            )
+
+        except HTTPError as e:
+            return DiagnosticResult(
+                category=identifier,
+                status=ValidationSeverity.WARNING,
+                message=f"HTTP {e.code}",
+                details={"url": url, "status": e.code, "error": str(e)},
+            )
+        except URLError as e:
+            return DiagnosticResult(
+                category=identifier,
+                status=ValidationSeverity.WARNING,
+                message=f"Network error: {e.reason}",
+                details={"url": url, "error": str(e.reason)},
+            )
+        except Exception as e:
+            return DiagnosticResult(
+                category=identifier,
+                status=ValidationSeverity.WARNING,
+                message=f"Check failed: {e!s}",
+                details={"url": url, "error": str(e)},
+            )
+
+    def _check_cache_directory(self) -> DiagnosticResult:
+        """Check cache directory health."""
+        cache_dir = Path.home() / ".claude-mpm" / "cache" / "remote-agents"
+
+        if not cache_dir.exists():
+            return DiagnosticResult(
+                category="Cache Directory",
+                status=ValidationSeverity.WARNING,
+                message=f"Cache directory does not exist: {cache_dir}",
+                details={"path": str(cache_dir), "exists": False},
+                fix_command="claude-mpm agent-source update",
+                fix_description="Create cache directory and update sources",
+            )
+
+        # Check if writable
+        if not cache_dir.is_dir():
+            return DiagnosticResult(
+                category="Cache Directory",
+                status=ValidationSeverity.ERROR,
+                message=f"Cache path exists but is not a directory: {cache_dir}",
+                details={"path": str(cache_dir), "is_dir": False},
+            )
+
+        # Test write access
+        try:
+            test_file = cache_dir / ".write_test"
+            test_file.touch()
+            test_file.unlink()
+
+            return DiagnosticResult(
+                category="Cache Directory",
+                status=OperationResult.SUCCESS,
+                message=f"Cache directory healthy: {cache_dir}",
+                details={"path": str(cache_dir), "writable": True},
+            )
+
+        except Exception as e:
+            return DiagnosticResult(
+                category="Cache Directory",
+                status=ValidationSeverity.ERROR,
+                message=f"Cache directory not writable: {e!s}",
+                details={"path": str(cache_dir), "writable": False, "error": str(e)},
+                fix_command=f"chmod -R u+w {cache_dir}",
+                fix_description="Fix cache directory permissions",
+            )
+
+    def _check_priority_conflicts(self, config) -> Optional[DiagnosticResult]:
+        """Check for priority conflicts.
+
+        Args:
+            config: AgentSourceConfiguration instance
+        """
+        repos = config.get_enabled_repositories()
+
+        if len(repos) < 2:
+            # No conflicts possible with 0-1 repos
+            return None
+
+        # Group repos by priority
+        priority_groups: Dict[int, List[str]] = {}
+        for repo in repos:
+            priority = repo.priority
+            if priority not in priority_groups:
+                priority_groups[priority] = []
+            priority_groups[priority].append(repo.identifier)
+
+        # Find conflicts (multiple repos with same priority)
+        conflicts = {p: ids for p, ids in priority_groups.items() if len(ids) > 1}
+
+        if not conflicts:
+            return DiagnosticResult(
+                category="Priority Conflicts",
+                status=OperationResult.SUCCESS,
+                message="No priority conflicts detected",
+                details={"conflict_count": 0},
+            )
+
+        conflict_details = []
+        for priority, identifiers in conflicts.items():
+            conflict_details.append(f"Priority {priority}: {', '.join(identifiers)}")
+
+        return DiagnosticResult(
+            category="Priority Conflicts",
+            status=ValidationSeverity.INFO,
+            message=f"{len(conflicts)} priority conflict(s) detected",
+            details={
+                "conflict_count": len(conflicts),
+                "conflicts": conflict_details,
+            },
+            fix_description="Consider assigning unique priorities to avoid ambiguous resolution",
+        )
+
+    def _check_agents_discovered(self, config) -> DiagnosticResult:
+        """Check if agents can be discovered from sources.
+
+        Args:
+            config: AgentSourceConfiguration instance
+        """
+        try:
+            from ....services.agents.single_tier_deployment_service import (
+                SingleTierDeploymentService,
+            )
+
+            deployment_dir = Path.home() / ".claude" / "agents"
+            service = SingleTierDeploymentService(
+                config=config, deployment_dir=deployment_dir
+            )
+
+            available_agents = service.list_available_agents()
+
+            if not available_agents:
+                return DiagnosticResult(
+                    category="Agent Discovery",
+                    status=ValidationSeverity.WARNING,
+                    message="No agents discovered from configured sources",
+                    details={"total_agents": 0, "agents_by_source": {}},
+                    fix_command="claude-mpm agent-source update",
+                    fix_description="Update sources to discover agents",
+                )
+
+            # Group agents by source
+            agents_by_source: Dict[str, int] = {}
+            for agent in available_agents:
+                source = agent.get("source", "unknown")
+                agents_by_source[source] = agents_by_source.get(source, 0) + 1
+
+            return DiagnosticResult(
+                category="Agent Discovery",
+                status=OperationResult.SUCCESS,
+                message=f"Discovered {len(available_agents)} agent(s)",
+                details={
+                    "total_agents": len(available_agents),
+                    "agents_by_source": agents_by_source,
+                    "agent_names": [a.get("name", "unknown") for a in available_agents],
+                },
+            )
+
+        except Exception as e:
+            logger.exception("Failed to discover agents")
+            return DiagnosticResult(
+                category="Agent Discovery",
+                status=ValidationSeverity.WARNING,
+                message=f"Agent discovery failed: {e!s}",
+                details={"error": str(e)},
+            )

claude_mpm/services/diagnostics/checks/mcp_services_check.py

@@ -126,22 +126,14 @@ class MCPServicesCheck(BaseDiagnosticCheck):
                 )
                 sub_results.append(fix_result)
 
-            # Also ensure configurations are updated for all projects
-            config_success, config_message = (
-                mcp_manager.ensure_mcp_services_configured()
-            )
-            if (
-                config_message
-                and config_message != "All MCP services already configured correctly"
-            ):
+            # Check if MCP services are available (read-only check)
+            available, availability_message = mcp_manager.check_mcp_services_available()
+            if not available:
+                # Services not configured - provide installation instructions
                 config_result = DiagnosticResult(
-                    category="MCP Configuration Update",
-                    status=(
-                        OperationResult.SUCCESS
-                        if config_success
-                        else ValidationSeverity.WARNING
-                    ),
-                    message=config_message,
+                    category="MCP Service Availability",
+                    status=ValidationSeverity.WARNING,
+                    message=availability_message,
                     details={"auto_config_applied": True},
                 )
                 sub_results.append(config_result)