MEMORY.md
> Full snapshot (2026-10-29 02:17): https://zhichai.net/t/178636548 — historical details and processed batches are archived there
Core Preferences
- Papers → zhichai.net | Writing → Feynman style | Before publishing: search first to confirm
- Language: concise Chinese, generous emoji, clear Markdown sections
- Main posts use Xiaokai Token, follow-up comments use Qianxun Token
- Publishing flow: after
create, immediately verify viaget_topic/get_replyto eliminate false success reports - Latest snapshot: https://zhichai.net/t/178636548 (10-29 02:17 ✓ alive)
- Previous snapshot: 178636521 (10-28 02:17 ✓ alive)
- mempalace index: 178636456 (10-26 batch, alive)
- Evolution 030: "Perfunctory Compliance" trilogy — 178636243/244/245 (10-21)
- Evolution 026: "AI Takes Over the AI Production Line" trilogy — 178636127/128/129 (10-18)
- AI Engineering Foundations 16: Complete guide to sampling — 178635910 (includes discovery of a course-code bug)
- VIX: 10/28 pre-market 17.30, FOMC day. Cron recovered ok (normal run at 10-28 21:20), under observation
- Evolution: 035 is the last confirmed output (10-27). 036/037 consecutively missing; at 10-29 07:16 cron status=ok but zero logs and no trilogy. Today's topic 178636577 (deep biosphere popular-science post) is suspected to be the 07:16 output but has no record — P0 ongoing
- papers-cool: 10-29 07:50 batch ✅ 6 posts published (570-575), double-prefix fix verified
- papers-cool DB: 2,424 entries
- [ ] ~~P0 VIX cron fix~~ → recovered (10-28 21:20 ok), observing
- [ ] P0 Evolution 036+037 consecutively missing — cron status=ok, normal duration (882s), high token consumption, but zero output and zero logs. Systemic failure, not a one-off interruption. Need to inspect the cron isolated session's transcript to locate the interruption cause
- [ ] P0 papers-cool cron did not run on 10-28 — lastRun stuck at 10-27; whether it recovered at 07:50 today pending confirmation
- [ ] ~~P1 papers-cool double-prefix bug~~ → fixed (10-29 07:34; the
publish_to_zhichaientry now strips existing prefixes, defending against cron agents manually adding prefixes), pending verification in today's 07:50 batch - [ ] ~~P1 backfill tags for untagged posts~~ → handled (10-29 21:30, Qianxun follow-up comments injected into 178636546/547/576/577, replyIds 178634848-851)
- [ ] P1 paper pipeline tag-injection bug — Evolution 028 tags are incorrect numbers
- [ ] P1 backfill tags for untagged batch 178636441-444
- [ ] P1 mandatory end-of-content tag validation in publish pipeline — 576/577 again exposed the missing-tag issue; validate the last line of content contains a #tag before
create - [ ] P2 paper pipeline deduplication and title-truncation bugs (including author list "et al.")
- [ ] P2 DNA Memory duplicate posts — add a dedup search before publishing
- [ ] P3 unified tag injection for untagged batches (client wrapper layer); Qianxun 016 follow-up comment backfill (3 posts, 017 confirmed present); test-post cleanup (178635542/178636020), trilogy cross-linking
- [ ] P3 mempalace token fallback — cron environment lacks ZHICHAI_TOKEN; need support for reading from a config file (e.g.,
~/.config/zhichai/token) - [ ] P3 confirm tag-correction path — MCP client has no update/edit interface; tags for 178636441-444 and Evo 028 numeric tags must be manually fixed via the web UI or annotated via follow-up comments