claude-mpm 3.4.10__py3-none-any.whl → 5.4.85__py3-none-any.whl

claude_mpm/agents/frontmatter_validator.py

@@ -0,0 +1,754 @@
+#!/usr/bin/env python3
+"""
+Frontmatter Validator for Agent Files
+=====================================
+
+This module provides validation and automatic correction for agent file frontmatter
+in .md, .claude, and .claude-mpm files. It ensures consistency and compatibility
+across different agent file formats.
+
+Key Features:
+- Validates frontmatter against a defined schema
+- Automatically corrects common formatting issues
+- Normalizes model names to standard tiers (opus, sonnet, haiku)
+- Fixes tools field when provided as string representation
+- Provides detailed logging of corrections made
+"""
+
+import json
+import re
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any, Dict, List, Optional, Tuple
+
+import yaml
+
+from claude_mpm.core.enums import ModelTier
+from claude_mpm.core.logging_utils import get_logger
+
+logger = get_logger(__name__)
+
+
+@dataclass
+class ValidationResult:
+    """Result of frontmatter validation."""
+
+    is_valid: bool
+    errors: List[str]
+    warnings: List[str]
+    corrections: List[str]
+    corrected_frontmatter: Optional[Dict[str, Any]] = None
+    field_corrections: Optional[Dict[str, Any]] = (
+        None  # Specific field-level corrections
+    )
+
+
+class FrontmatterValidator:
+    """
+    Validates and corrects frontmatter in agent files.
+
+    This class handles:
+    - Schema validation against frontmatter_schema.json
+    - Automatic correction of common issues
+    - Model name normalization
+    - Tools field parsing and correction
+    - Logging of all corrections made
+    """
+
+    # NOTE: Model normalization now handled by ModelTier.normalize()
+    # This enum-based approach replaced 26 lines of manual mappings
+
+    # Tool name corrections (case normalization)
+    TOOL_CORRECTIONS = {
+        "read": "Read",
+        "write": "Write",
+        "edit": "Edit",
+        "multiedit": "MultiEdit",
+        "grep": "Grep",
+        "glob": "Glob",
+        "ls": "LS",
+        "bash": "Bash",
+        "websearch": "WebSearch",
+        "webfetch": "WebFetch",
+        "notebookread": "NotebookRead",
+        "notebookedit": "NotebookEdit",
+        "todowrite": "TodoWrite",
+        "exitplanmode": "ExitPlanMode",
+    }
+
+    # Valid tool names
+    VALID_TOOLS = {
+        "Read",
+        "Write",
+        "Edit",
+        "MultiEdit",
+        "Grep",
+        "Glob",
+        "LS",
+        "Bash",
+        "WebSearch",
+        "WebFetch",
+        "NotebookRead",
+        "NotebookEdit",
+        "TodoWrite",
+        "ExitPlanMode",
+        "git",
+        "docker",
+        "kubectl",
+        "terraform",
+        "aws",
+        "gcloud",
+        "azure",
+    }
+
+    # Valid model tiers
+    VALID_MODELS = {"opus", "sonnet", "haiku"}
+
+    def __init__(self):
+        """Initialize the validator with schema if available."""
+        self.schema = self._load_schema()
+        self.all_valid_fields = self._extract_valid_fields()
+
+    def _load_schema(self) -> Optional[Dict[str, Any]]:
+        """Load the frontmatter schema from JSON file."""
+        schema_path = (
+            Path(__file__).parent.parent / "schemas" / "frontmatter_schema.json"
+        )
+        if schema_path.exists():
+            try:
+                with schema_path.open() as f:
+                    return json.load(f)
+            except Exception as e:
+                logger.warning(f"Failed to load frontmatter schema: {e}")
+        return None
+
+    def _extract_valid_fields(self) -> set:
+        """Extract all valid field names from the schema."""
+        if self.schema and "properties" in self.schema:
+            return set(self.schema["properties"].keys())
+        # Fallback to known fields if schema not available
+        return {
+            "name",
+            "description",
+            "version",
+            "base_version",
+            "author",
+            "tools",
+            "model",
+            "tags",
+            "category",
+            "max_tokens",
+            "temperature",
+            "resource_tier",
+            "dependencies",
+            "capabilities",
+            "color",
+            # NEW: Collection-based identification fields
+            "collection_id",
+            "source_path",
+            "canonical_id",
+        }
+
+    def validate_and_correct(self, frontmatter: Dict[str, Any]) -> ValidationResult:
+        """
+        Validate and automatically correct frontmatter.
+
+        Args:
+            frontmatter: Dictionary of frontmatter fields
+
+        Returns:
+            ValidationResult with validation status and corrected frontmatter
+        """
+        errors: List[str] = []
+        warnings: List[str] = []
+        corrections: List[str] = []
+        corrected = frontmatter.copy()
+        field_corrections: Dict[str, Any] = {}
+
+        # Check required fields
+        self._validate_required_fields(corrected, errors)
+
+        # Validate and correct individual fields
+        self._validate_name_field(corrected, field_corrections, errors, corrections)
+        self._validate_model_field(corrected, field_corrections, errors, corrections)
+        self._validate_tools_field(corrected, field_corrections, warnings, corrections)
+        self._validate_version_fields(corrected, field_corrections, errors, corrections)
+        self._validate_description_field(corrected, errors, warnings)
+        self._validate_category_field(corrected, warnings)
+        self._validate_resource_tier_field(corrected, warnings)
+        self._validate_color_field(corrected, errors)
+        self._validate_author_field(corrected, errors, warnings)
+        self._validate_tags_field(corrected, errors, warnings)
+        self._validate_numeric_fields(corrected, errors, warnings)
+        # NEW: Validate collection-based identification fields
+        self._validate_collection_fields(corrected, field_corrections, errors, warnings)
+
+        # Determine if valid
+        is_valid = len(errors) == 0
+
+        return ValidationResult(
+            is_valid=is_valid,
+            errors=errors,
+            warnings=warnings,
+            corrections=corrections,
+            corrected_frontmatter=corrected if corrections else None,
+            field_corrections=field_corrections if field_corrections else None,
+        )
+
+    def _validate_required_fields(
+        self, corrected: Dict[str, Any], errors: List[str]
+    ) -> None:
+        """Check that all required fields are present."""
+        required_fields = (
+            self.schema.get("required", ["name", "description", "version"])
+            if self.schema
+            else ["name", "description", "version"]
+        )
+        for field in required_fields:
+            if field not in corrected:
+                errors.append(f"Missing required field: {field}")
+
+    def _validate_name_field(
+        self,
+        corrected: Dict[str, Any],
+        field_corrections: Dict[str, Any],
+        errors: List[str],
+        corrections: List[str],
+    ) -> None:
+        """Validate and correct the name field."""
+        if "name" not in corrected:
+            return
+
+        name = corrected["name"]
+        if not isinstance(name, str):
+            errors.append(f"Field 'name' must be a string, got {type(name).__name__}")
+            return
+
+        if not re.match(r"^[a-z][a-z0-9_]*$", name):
+            # Try to fix the name
+            fixed_name = name.lower().replace("-", "_").replace(" ", "_")
+            fixed_name = re.sub(r"[^a-z0-9_]", "", fixed_name)
+            if fixed_name and fixed_name[0].isalpha():
+                corrected["name"] = fixed_name
+                field_corrections["name"] = fixed_name
+                corrections.append(f"Corrected name from '{name}' to '{fixed_name}'")
+            else:
+                errors.append(f"Invalid name format: {name}")
+
+    def _validate_model_field(
+        self,
+        corrected: Dict[str, Any],
+        field_corrections: Dict[str, Any],
+        errors: List[str],
+        corrections: List[str],
+    ) -> None:
+        """Validate and correct the model field."""
+        if "model" not in corrected:
+            return
+
+        model = corrected["model"]
+
+        # Convert to string if it's a number (YAML might parse dates as integers)
+        if isinstance(model, (int, float)):
+            model = str(model)
+            corrected["model"] = model
+            field_corrections["model"] = model
+            corrections.append(f"Converted model from number to string: {model}")
+
+        if not isinstance(model, str):
+            errors.append(f"Field 'model' must be a string, got {type(model).__name__}")
+            return
+
+        normalized_model = self._normalize_model(model)
+        if normalized_model != model:
+            corrected["model"] = normalized_model
+            field_corrections["model"] = normalized_model
+            corrections.append(
+                f"Normalized model from '{model}' to '{normalized_model}'"
+            )
+
+        if normalized_model not in self.VALID_MODELS:
+            errors.append(f"Invalid model: {model} (normalized to {normalized_model})")
+
+    def _validate_tools_field(
+        self,
+        corrected: Dict[str, Any],
+        field_corrections: Dict[str, Any],
+        warnings: List[str],
+        corrections: List[str],
+    ) -> None:
+        """Validate and correct the tools field."""
+        if "tools" not in corrected:
+            return
+
+        tools = corrected["tools"]
+        corrected_tools, tool_corrections = self._correct_tools(tools)
+        if tool_corrections:
+            corrected["tools"] = corrected_tools
+            field_corrections["tools"] = corrected_tools
+            corrections.extend(tool_corrections)
+
+        # Validate tool names
+        invalid_tools = []
+        for tool in corrected_tools:
+            if tool not in self.VALID_TOOLS:
+                # Try to correct the tool name
+                corrected_tool = self.TOOL_CORRECTIONS.get(tool.lower())
+                if corrected_tool:
+                    idx = corrected_tools.index(tool)
+                    corrected_tools[idx] = corrected_tool
+                    corrected["tools"] = corrected_tools
+                    field_corrections["tools"] = corrected_tools
+                    corrections.append(f"Corrected tool '{tool}' to '{corrected_tool}'")
+                else:
+                    invalid_tools.append(tool)
+
+        if invalid_tools:
+            warnings.append(f"Unknown tools: {', '.join(invalid_tools)}")
+
+    def _validate_version_fields(
+        self,
+        corrected: Dict[str, Any],
+        field_corrections: Dict[str, Any],
+        errors: List[str],
+        corrections: List[str],
+    ) -> None:
+        """Validate and correct version fields."""
+        version_fields = ["version", "base_version"]
+        for field in version_fields:
+            if field not in corrected:
+                continue
+
+            version = corrected[field]
+            if not isinstance(version, str):
+                errors.append(
+                    f"Field '{field}' must be a string, got {type(version).__name__}"
+                )
+                continue
+
+            if re.match(r"^\d+\.\d+\.\d+$", version):
+                continue  # Valid format
+
+            # Try to fix common version issues
+            if re.match(r"^\d+\.\d+$", version):
+                fixed_version = f"{version}.0"
+                corrected[field] = fixed_version
+                field_corrections[field] = fixed_version
+                corrections.append(
+                    f"Fixed {field} from '{version}' to '{fixed_version}'"
+                )
+            elif re.match(r"^v?\d+\.\d+\.\d+$", version):
+                fixed_version = version.lstrip("v")
+                corrected[field] = fixed_version
+                field_corrections[field] = fixed_version
+                corrections.append(
+                    f"Fixed {field} from '{version}' to '{fixed_version}'"
+                )
+            else:
+                errors.append(f"Invalid {field} format: {version}")
+
+    def _validate_description_field(
+        self, corrected: Dict[str, Any], errors: List[str], warnings: List[str]
+    ) -> None:
+        """Validate the description field."""
+        if "description" not in corrected:
+            return
+
+        desc = corrected["description"]
+        if not isinstance(desc, str):
+            errors.append(
+                f"Field 'description' must be a string, got {type(desc).__name__}"
+            )
+        elif len(desc) < 10:
+            warnings.append(f"Description too short ({len(desc)} chars, minimum 10)")
+        elif len(desc) > 200:
+            warnings.append(f"Description too long ({len(desc)} chars, maximum 200)")
+
+    def _validate_category_field(
+        self, corrected: Dict[str, Any], warnings: List[str]
+    ) -> None:
+        """Validate the category field."""
+        if "category" not in corrected:
+            return
+
+        valid_categories = [
+            "engineering",
+            "research",
+            "quality",
+            "operations",
+            "specialized",
+        ]
+        if corrected["category"] not in valid_categories:
+            warnings.append(f"Invalid category: {corrected['category']}")
+
+    def _validate_resource_tier_field(
+        self, corrected: Dict[str, Any], warnings: List[str]
+    ) -> None:
+        """Validate the resource_tier field."""
+        if "resource_tier" not in corrected:
+            return
+
+        valid_tiers = ["basic", "standard", "intensive", "lightweight"]
+        if corrected["resource_tier"] not in valid_tiers:
+            warnings.append(f"Invalid resource_tier: {corrected['resource_tier']}")
+
+    def _validate_color_field(
+        self, corrected: Dict[str, Any], errors: List[str]
+    ) -> None:
+        """Validate the color field."""
+        if "color" not in corrected:
+            return
+
+        color = corrected["color"]
+        if not isinstance(color, str):
+            errors.append(f"Field 'color' must be a string, got {type(color).__name__}")
+
+    def _validate_author_field(
+        self, corrected: Dict[str, Any], errors: List[str], warnings: List[str]
+    ) -> None:
+        """Validate the author field."""
+        if "author" not in corrected:
+            return
+
+        author = corrected["author"]
+        if not isinstance(author, str):
+            errors.append(
+                f"Field 'author' must be a string, got {type(author).__name__}"
+            )
+        elif len(author) > 100:
+            warnings.append(f"Author field too long ({len(author)} chars, maximum 100)")
+
+    def _validate_tags_field(
+        self, corrected: Dict[str, Any], errors: List[str], warnings: List[str]
+    ) -> None:
+        """Validate the tags field."""
+        if "tags" not in corrected:
+            return
+
+        tags = corrected["tags"]
+        if isinstance(tags, str):
+            # Convert comma-separated string to list for validation
+            tag_list = [tag.strip() for tag in tags.split(",") if tag.strip()]
+        elif isinstance(tags, list):
+            tag_list = tags
+        else:
+            errors.append(
+                f"Field 'tags' must be a list or comma-separated string, got {type(tags).__name__}"
+            )
+            return
+
+        for tag in tag_list:
+            if not isinstance(tag, str):
+                errors.append(f"All tags must be strings, found {type(tag).__name__}")
+            elif not re.match(r"^[a-z][a-z0-9-]*$", tag):
+                warnings.append(
+                    f"Tag '{tag}' doesn't match recommended pattern (lowercase, alphanumeric with hyphens)"
+                )
+
+    def _validate_numeric_fields(
+        self, corrected: Dict[str, Any], errors: List[str], warnings: List[str]
+    ) -> None:
+        """Validate numeric fields (max_tokens, temperature)."""
+        for field_name, (min_val, max_val) in [
+            ("max_tokens", (1000, 200000)),
+            ("temperature", (0, 1)),
+        ]:
+            if field_name not in corrected:
+                continue
+
+            value = corrected[field_name]
+            if field_name == "temperature" and not isinstance(value, (int, float)):
+                errors.append(
+                    f"Field '{field_name}' must be a number, got {type(value).__name__}"
+                )
+            elif field_name == "max_tokens" and not isinstance(value, int):
+                errors.append(
+                    f"Field '{field_name}' must be an integer, got {type(value).__name__}"
+                )
+            elif isinstance(value, (int, float)) and not (min_val <= value <= max_val):
+                warnings.append(
+                    f"Field '{field_name}' value {value} outside recommended range [{min_val}, {max_val}]"
+                )
+
+    def _validate_collection_fields(
+        self,
+        corrected: Dict[str, Any],
+        field_corrections: Dict[str, Any],
+        errors: List[str],
+        warnings: List[str],
+    ) -> None:
+        """Validate collection-based identification fields.
+
+        NEW: Validates collection_id, source_path, and canonical_id fields.
+
+        These fields are auto-populated by RemoteAgentDiscoveryService for remote agents
+        and should follow specific formats:
+        - collection_id: "owner/repo-name" (e.g., "bobmatnyc/claude-mpm-agents")
+        - source_path: Relative path in repo (e.g., "agents/pm.md")
+        - canonical_id: "collection_id:agent_id" or "legacy:filename"
+        """
+        # Validate collection_id format (optional field)
+        if "collection_id" in corrected:
+            collection_id = corrected["collection_id"]
+            if not isinstance(collection_id, str):
+                errors.append(
+                    f"Field 'collection_id' must be a string, got {type(collection_id).__name__}"
+                )
+            elif "/" not in collection_id:
+                warnings.append(
+                    f"Field 'collection_id' should be in format 'owner/repo-name', got '{collection_id}'"
+                )
+
+        # Validate source_path format (optional field)
+        if "source_path" in corrected:
+            source_path = corrected["source_path"]
+            if not isinstance(source_path, str):
+                errors.append(
+                    f"Field 'source_path' must be a string, got {type(source_path).__name__}"
+                )
+
+        # Validate canonical_id format (optional field)
+        if "canonical_id" in corrected:
+            canonical_id = corrected["canonical_id"]
+            if not isinstance(canonical_id, str):
+                errors.append(
+                    f"Field 'canonical_id' must be a string, got {type(canonical_id).__name__}"
+                )
+            elif ":" not in canonical_id:
+                warnings.append(
+                    f"Field 'canonical_id' should be in format 'collection:agent_id' or 'legacy:filename', got '{canonical_id}'"
+                )
+
+        # Auto-generate canonical_id if collection_id is present but canonical_id is missing
+        if "collection_id" in corrected and "canonical_id" not in corrected:
+            collection_id = corrected["collection_id"]
+            agent_id = corrected.get("name", "unknown")
+
+            # Generate canonical_id
+            canonical_id = f"{collection_id}:{agent_id}"
+            corrected["canonical_id"] = canonical_id
+            field_corrections["canonical_id"] = canonical_id
+            warnings.append(
+                f"Auto-generated canonical_id: '{canonical_id}' from collection_id and name"
+            )
+
+    def _normalize_model(self, model: str) -> str:
+        """
+        Normalize model name to standard tier using ModelTier enum.
+
+        Args:
+            model: Original model name
+
+        Returns:
+            Normalized model tier name
+        """
+        return ModelTier.normalize(model).value
+
+    def _correct_tools(self, tools: Any) -> Tuple[List[str], List[str]]:
+        """
+        Correct tools field formatting issues.
+
+        Args:
+            tools: Original tools value (could be string, list, etc.)
+
+        Returns:
+            Tuple of (corrected_tools_list, list_of_corrections)
+        """
+        corrections = []
+
+        # If already a proper list, just validate
+        if isinstance(tools, list):
+            return tools, corrections
+
+        # If it's a string, try to parse it
+        if isinstance(tools, str):
+            # Remove any surrounding whitespace
+            tools_str = tools.strip()
+
+            # Check if it's a string representation of a list
+            if tools_str.startswith("[") and tools_str.endswith("]"):
+                # Try to parse as JSON array
+                try:
+                    parsed_tools = json.loads(tools_str)
+                    if isinstance(parsed_tools, list):
+                        corrections.append(
+                            f"Parsed tools from JSON string: {tools_str[:50]}..."
+                        )
+                        return parsed_tools, corrections
+                except json.JSONDecodeError:
+                    pass
+
+                # Try to extract comma-separated values
+                tools_str = tools_str[1:-1]  # Remove brackets
+
+            # Split by comma and clean up
+            if "," in tools_str:
+                tool_list = [t.strip().strip("'\"") for t in tools_str.split(",")]
+            else:
+                # Single tool or space-separated
+                tool_list = tools_str.replace(",", " ").split()
+
+            # Clean up tool names
+            cleaned_tools = []
+            for tool in tool_list:
+                tool = tool.strip().strip("'\"")
+                if tool:
+                    cleaned_tools.append(tool)
+
+            if cleaned_tools:
+                corrections.append(
+                    f"Converted tools from string to list: {len(cleaned_tools)} tools"
+                )
+                return cleaned_tools, corrections
+
+        # If we can't parse it, return empty list
+        corrections.append(
+            f"Could not parse tools field (type: {type(tools).__name__}), using empty list"
+        )
+        return [], corrections
+
+    def validate_file(self, file_path: Path) -> ValidationResult:
+        """
+        Validate frontmatter in a specific file.
+
+        Args:
+            file_path: Path to agent file
+
+        Returns:
+            ValidationResult with validation status
+        """
+        try:
+            with file_path.open() as f:
+                content = f.read()
+
+            # Extract frontmatter
+            frontmatter = self._extract_frontmatter(content)
+            if not frontmatter:
+                return ValidationResult(
+                    is_valid=False,
+                    errors=["No frontmatter found in file"],
+                    warnings=[],
+                    corrections=[],
+                )
+
+            # Validate and correct
+            return self.validate_and_correct(frontmatter)
+
+        except Exception as e:
+            return ValidationResult(
+                is_valid=False,
+                errors=[f"Error reading file: {e}"],
+                warnings=[],
+                corrections=[],
+            )
+
+    def _extract_frontmatter(self, content: str) -> Optional[Dict[str, Any]]:
+        """
+        Extract frontmatter from file content.
+
+        Args:
+            content: File content
+
+        Returns:
+            Parsed frontmatter dictionary or None
+        """
+        # Check for YAML frontmatter (between --- markers)
+        if content.startswith("---"):
+            try:
+                end_marker = content.find("\n---\n", 4)
+                if end_marker == -1:
+                    end_marker = content.find("\n---\r\n", 4)
+
+                if end_marker != -1:
+                    frontmatter_str = content[4:end_marker]
+                    return yaml.safe_load(frontmatter_str)
+            except yaml.YAMLError as e:
+                logger.error(f"Failed to parse YAML frontmatter: {e}")
+
+        return None
+
+    def correct_file(self, file_path: Path, dry_run: bool = False) -> ValidationResult:
+        """
+        Validate and optionally correct a file's frontmatter.
+
+        Args:
+            file_path: Path to agent file
+            dry_run: If True, don't write changes to file
+
+        Returns:
+            ValidationResult with corrections made
+        """
+        result = self.validate_file(file_path)
+
+        if result.field_corrections and not dry_run:
+            try:
+                with file_path.open() as f:
+                    content = f.read()
+
+                # Find frontmatter boundaries
+                if content.startswith("---"):
+                    end_marker = content.find("\n---\n", 4)
+                    if end_marker == -1:
+                        end_marker = content.find("\n---\r\n", 4)
+
+                    if end_marker != -1:
+                        # Apply field-level corrections to preserve structure
+                        frontmatter_content = content[4:end_marker]
+                        corrected_content = self._apply_field_corrections(
+                            frontmatter_content, result.field_corrections
+                        )
+
+                        if corrected_content != frontmatter_content:
+                            new_content = f"---\n{corrected_content}\n---\n{content[end_marker + 5 :]}"
+
+                            with file_path.open("w") as f:
+                                f.write(new_content)
+
+                            logger.info(f"Corrected frontmatter in {file_path}")
+                            for correction in result.corrections:
+                                logger.info(f"  - {correction}")
+            except Exception as e:
+                logger.error(f"Failed to write corrections to {file_path}: {e}")
+
+        return result
+
+    def _apply_field_corrections(
+        self, frontmatter_content: str, field_corrections: Dict[str, Any]
+    ) -> str:
+        """
+        Apply field-level corrections while preserving structure and other fields.
+
+        Args:
+            frontmatter_content: Original YAML frontmatter content
+            field_corrections: Dict of field corrections to apply
+
+        Returns:
+            Corrected frontmatter content
+        """
+        lines = frontmatter_content.strip().split("\n")
+        corrected_lines = []
+
+        for line in lines:
+            # Check if this line contains a field we need to correct
+            if ":" in line:
+                field_name = line.split(":")[0].strip()
+                if field_name in field_corrections:
+                    # Replace the field value while preserving structure
+                    corrected_value = field_corrections[field_name]
+                    if isinstance(corrected_value, list):
+                        # Handle list fields like tools
+                        if field_name == "tools" and isinstance(corrected_value, list):
+                            # Format as comma-separated string to preserve existing format
+                            corrected_lines.append(
+                                f"{field_name}: {','.join(corrected_value)}"
+                            )
+                        else:
+                            corrected_lines.append(f"{field_name}: {corrected_value}")
+                    else:
+                        corrected_lines.append(f"{field_name}: {corrected_value}")
+                    continue
+
+            # Keep the original line if no correction needed
+            corrected_lines.append(line)
+
+        return "\n".join(corrected_lines)