Explorer B assignment

Read completely kb/work/open-ended-work-economics/curiosity-exploration/README.md, kb/work/open-ended-work-economics/bootstrap-exploration/second-exploration-brief.md, kb/work/open-ended-work-economics/README.md, AGENTS.md and kb/work/COLLECTION.md. These bind purpose, scope, evidence, coordination, acceptance and stop conditions. No code, experiments, implementation, commits, extraction, deletion or library/index edits. No new source destination allocated; return ungrounded literature leads as leads, not source-backed conclusions. Existing grounded Quotes may support claims; source analysis is not support. Only A turn 1 had source mutation ownership. Choose methods and research judgments within scope; escalate purpose/authority changes, return partial on missing evidence/tools. Validate owned result. Distinguish source claims, documentary observations, conjectures and demonstrations; preserve strongest rival, discriminating evidence, recoverability needs and reason for parked alternatives. Keep learning value distinct from task benefit; do not force compression scores, text shortening or service-first admission. Fresh Explorer B turn 4. Read prior integrations and current coordination, not current A. Independently investigate acquisition, selective retention, reacquisition and hidden deterioration using concrete KB cases other than solely confidence. Distinguish exact reconstruction of acquired material from preserving ability to ask unforeseen questions. How could a later worker discover a discarded detail mattered; when is this impossible? How can old/new common-case checks conceal deterioration elsewhere? Propose modest practical acquisition/retention/consultation requirements, not a store/schema or keep-everything rule; identify competing costs and irrecoverable uncertainty. Keep question origination and intrinsic learning value central. Own only kb/work/open-ended-work-economics/curiosity-exploration/turn-04/explorer-b.md, 1200-1800 words. No new source destinations or mutations.