Skip to main content

Validation - MCP Tool Specification

This document specifies the MCP tools exposed by the Validation module via @mcp_tool decorators in mcp_tools.py. These tools are auto-discovered by the PAI MCP bridge.

Available MCP Tools​

validate_schema​

Validate arbitrary data against a JSON Schema or Pydantic model.

Parameters:

  • data (dict, required): Data to validate
  • schema (dict, required): JSON Schema definition or Pydantic model reference
  • validator_type (str, default "json_schema"): Strategy — "json_schema", "pydantic", or "custom"

Returns: {is_valid, errors: [{field, message}], warnings: [...]}

Trust level: Safe

validate_config​

Validate a configuration dictionary for required keys and type correctness.

Parameters:

  • config (dict, required): Configuration dictionary to validate
  • required_keys (list[str], optional): Keys that must be present
  • strict (bool, default false): Reject unknown keys if true

Returns: {is_valid, missing_keys: [...], extra_keys: [...], errors: [...]}

Trust level: Safe

validation_summary​

Return aggregate statistics from the validation manager.

Parameters: None

Returns: {run_count, pass_rate, error_rate, validators_used: [...]}

Trust level: Safe

For programmatic Python integration, refer to README.md and API_SPECIFICATION.md.