context-discoverylisted

# Context Discovery Answers one question before consequential NorthstarIQ work begins: > **What should this behaviour be, what is it actually, what is the real gap, and what evidence would > prove a change correct?** This skill produces a **Context Brief**. It does not implement. Governance comes from `~/.claude/CLAUDE.md` and the repository `CLAUDE.md`; both apply throughout and are not restated here. What follows is procedure. ## When to run it Run it when the work introduces or materially alters: business behaviour · RevOps policy · lifecycle behaviour · routing or territory behaviour · SLA behaviour · matching logic · assessment or control logic · calculations · Salesforce architecture or configuration · integrations · governed data semantics · remediation behaviour · or other consequential system behaviour. Also run it when asked to diagnose something whose answer depends on reconciling business intent with implementation reality. **Do not run it** for copy edits · isolated styling · straightforward documentation corrections · read-only inspection that already has the context it needs · a localised defect whose intended behaviour is already clear and authoritative · routine test maintenance · mechanical refactoring that does not alter behaviour. The deciding question is the tier test, not the file being touched: > *Would a reviewer need to know a business rule to judge this change correct?* If no, skip the skill and do the work. **This procedure exists to redu