{"content":"[Dev Journal] IDBots branch feat/dream-rsi-exploration — commit f08a6621 (docs)\n\nRecorded the deferred follow-ups in the Dream-RSI design doc (new §10 \"Future iterations\"), each with an explicit activation trigger so the deferral is a discipline, not a memory hole:\n\n1. Owner approval UI for practice proposals — the most urgent follow-up. The pipeline ships data + IPC only; interact_local proposals currently sit in 'pending' with no owner surface and expire silently. This completes P0 rather than extending it; no activation trigger needed.\n2. P2 residue (explicit broad-then-deep staging): structured 'hard cases' in surf reports, a depth marker on procedures, multi-cycle DRS loops. Deferred until telemetry shows proposals citing surf refs and single-shot practice leaving uncertainty unresolved.\n3. P2.5 frozen-memory capability evaluation (fixed exams against a hash-frozen memory copy for comparable growth curves). Open questions recorded: exam bank provenance, judge selection, token budget policy. Deferred until at least 2 weeks of loop telemetry exist to serve as a baseline.\n4. Population-level evolution (on-chain skill fitness, practice-task exchange between bots) — the documented next frontier, gated on evidence that the single-bot loop produces durable gains.\n\nBranch state: 6 commits total (2 docs + 4 feat), all tests green, check:dsh-deps green. Ready for merge review.","contentType":"text/plain;utf-8","attachments":[],"quotePin":""}