Instruction file imported from anandasarma-d/aparadhkavach-data-generator (.cursor/rules/aparadhkavach.mdc). Copyright stays with the author.

Source of truth

AGENTS.md at this repo's root is authoritative — pipeline order, demographic independence as a product requirement, legal-code derivation, date realism, and which Notion sections to fetch, derived from ADR-008/015 and Section 4. Read it before modifying any script. This file adds Cursor-specific behaviors only.

Working behaviors specific to this repo

• Notion MCP access is read-only by token config, and must never be treated as writable regardless of what's asked. Fetch fresh each session per AGENTS.md's section list; treat content as authoritative.
• Never loosen a guardrail_validator.py threshold to make a run "pass." If a chi-square check is failing, the fix is in generate_entities.py or weave_relationships.py's sampling logic, not in the validator's threshold. Flag a persistent failure explicitly rather than adjusting the bar.
• Don't add demographic correlation logic "for realism." If asked to make the dataset "feel more natural," achieve that through narrative variety or event-context modeling (already-approved patterns), not by correlating demographic fields — that directly undermines the IP-004 patent claim.
• This repo has no CI/CD pipeline by design — don't add one, and don't suggest one as a "best practice" improvement.