JSON Data Quality Audit

Perform a deterministic first-pass audit before deeper analysis.

Instructions

1. Parse the payload as JSON; report parse failure without repairing silently.
2. Walk the structure and count objects, arrays, strings, numbers, booleans, nulls, keys, array items, and maximum depth.
3. For a top-level array, canonicalize each row consistently and identify exact duplicate rows by index.
4. Compute a SHA-256 digest of the canonical serialized payload when hashing is available.
5. Report observable schema inconsistencies without guessing intended values.
6. Keep remediation separate from evidence.

Output

Return: valid_json, structural_stats, exact_duplicate_count, duplicate_indices, canonical_sha256 when available, findings[], suggested_cleanup[].

Boundaries

Do not expose secrets. Do not infer that two non-identical rows represent the same person/entity without a supplied matching rule. Do not mutate the source unless asked.

Version

v1