{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://github.com/GRCEngClub/claude-grc-engineering/schemas/finding.schema.json",
  "title": "GRC Finding",
  "description": "Canonical output contract for claude-grc-engineering connectors. One document per resource-evaluation batch. See docs/ARCHITECTURE.md for the pipeline model.",
  "type": "object",
  "required": [
    "schema_version",
    "source",
    "source_version",
    "run_id",
    "collected_at",
    "resource",
    "evaluations"
  ],
  "additionalProperties": false,
  "properties": {
    "schema_version": {
      "type": "string",
      "description": "Semver of this schema. Consumers pin a major version.",
      "pattern": "^[0-9]+\\.[0-9]+\\.[0-9]+$",
      "const": "1.0.0"
    },
    "source": {
      "type": "string",
      "description": "Stable identifier of the connector that produced this document.",
      "examples": ["aws-sec-inspector", "okta-inspector-py", "github-sec-inspector", "gcp-sec-inspector"],
      "pattern": "^[a-z][a-z0-9-]*$"
    },
    "source_version": {
      "type": "string",
      "description": "Version of the connector tool, for reproducibility.",
      "examples": ["1.2.3", "2026.04.13"]
    },
    "run_id": {
      "type": "string",
      "description": "Unique identifier for this collection run. ULID or UUID recommended.",
      "minLength": 8
    },
    "collected_at": {
      "type": "string",
      "format": "date-time",
      "description": "ISO 8601 timestamp when collection completed."
    },
    "resource": {
      "type": "object",
      "description": "The resource that was evaluated.",
      "required": ["type", "id"],
      "additionalProperties": true,
      "properties": {
        "type": {
          "type": "string",
          "description": "Resource type, prefixed by provider. Use snake_case.",
          "examples": ["aws_iam_user", "aws_s3_bucket", "okta_authentication_policy", "github_repository", "gcp_iam_binding"]
        },
        "id": {
          "type": "string",
          "description": "Provider-native identifier."
        },
        "arn": {
          "type": ["string", "null"],
          "description": "AWS ARN or equivalent canonical identifier if applicable."
        },
        "uri": {
          "type": ["string", "null"],
          "description": "Non-AWS canonical URI (e.g., gcp self-link, okta URL, github html_url)."
        },
        "region": {
          "type": ["string", "null"],
          "description": "Region or location. Null for global resources."
        },
        "account_id": {
          "type": ["string", "null"],
          "description": "AWS account ID, GCP project ID, Okta org, GitHub org, etc."
        },
        "tags": {
          "type": "object",
          "description": "Resource tags/labels if available.",
          "additionalProperties": {"type": "string"}
        }
      }
    },
    "evaluations": {
      "type": "array",
      "description": "One or more control evaluations performed against this resource. At least one is required.",
      "minItems": 1,
      "items": {"$ref": "#/$defs/evaluation"}
    },
    "raw_attributes": {
      "type": "object",
      "description": "Optional: the tool's native attributes for this resource. Preserved as-is for audit trail. Consumers SHOULD treat as opaque.",
      "additionalProperties": true
    },
    "findings": {
      "type": "array",
      "description": "Optional: high-level narrative findings that may span multiple controls.",
      "items": {"$ref": "#/$defs/narrative_finding"}
    },
    "metadata": {
      "type": "object",
      "description": "Optional connector-specific context (config used, scan scope, rate limits observed, etc.).",
      "additionalProperties": true
    }
  },
  "$defs": {
    "evaluation": {
      "type": "object",
      "description": "One control evaluation. A single resource may have many evaluations (e.g., an IAM user evaluated against AC-2, AC-6, IA-2, IA-5).",
      "required": ["control_framework", "control_id", "status"],
      "additionalProperties": false,
      "properties": {
        "control_framework": {
          "type": "string",
          "description": "Framework identifier. Prefer SCF for portable indexing; others are acceptable and are mapped via crosswalk.",
          "examples": ["SCF", "NIST-800-53-r5", "NIST-800-171-r3", "SOC2-TSC-2017", "PCI-DSS-4.0", "ISO-27001-2022", "CIS-v8", "HIPAA", "FedRAMP-Moderate", "FedRAMP-High", "HITRUST-CSF-11", "CMMC-2.0"]
        },
        "control_id": {
          "type": "string",
          "description": "Framework-native control identifier (e.g., SCF 'GOV-01', NIST 'AC-2', SOC2 'CC6.1').",
          "minLength": 1
        },
        "status": {
          "type": "string",
          "enum": ["pass", "fail", "not_applicable", "inconclusive", "skipped"],
          "description": "pass: requirement met. fail: requirement not met. not_applicable: scoped out. inconclusive: tool could not determine (missing permission, API error). skipped: intentionally skipped by user config."
        },
        "severity": {
          "type": "string",
          "enum": ["critical", "high", "medium", "low", "info"],
          "description": "Independent of status. A failed low-severity control is not a passed one. Use 'info' for passing/NA evaluations."
        },
        "message": {
          "type": "string",
          "description": "Human-readable detail. Required for status=fail or status=inconclusive."
        },
        "remediation": {
          "type": ["object", "null"],
          "description": "Optional structured remediation guidance.",
          "additionalProperties": false,
          "properties": {
            "summary": {"type": "string"},
            "ref": {
              "type": "string",
              "description": "Reference to a remediation artifact: generated Terraform, scan-iac auto-fix, or external runbook URL."
            },
            "effort_hours": {"type": "number", "minimum": 0},
            "automation": {
              "type": "string",
              "enum": ["auto_fixable", "semi_automated", "manual", "design_change"]
            }
          }
        },
        "evidence_refs": {
          "type": "array",
          "description": "Pointers to raw artifacts backing this evaluation. Path relative to the cache dir, or a URL.",
          "items": {"type": "string"}
        },
        "assessed_at": {
          "type": "string",
          "format": "date-time",
          "description": "Optional per-evaluation timestamp if different from collected_at."
        }
      },
      "allOf": [
        {
          "if": {"properties": {"status": {"const": "fail"}}},
          "then": {"required": ["message", "severity"]}
        },
        {
          "if": {"properties": {"status": {"const": "inconclusive"}}},
          "then": {"required": ["message"]}
        }
      ]
    },
    "narrative_finding": {
      "type": "object",
      "description": "High-level finding that may span multiple controls or resources.",
      "required": ["id", "title", "severity"],
      "additionalProperties": false,
      "properties": {
        "id": {"type": "string"},
        "title": {"type": "string"},
        "severity": {
          "type": "string",
          "enum": ["critical", "high", "medium", "low", "info"]
        },
        "description": {"type": "string"},
        "related_control_ids": {
          "type": "array",
          "items": {"type": "string"}
        },
        "related_resource_ids": {
          "type": "array",
          "items": {"type": "string"}
        }
      }
    }
  }
}