fix: persist mined writing patterns across skills

README.ja-JP.md

@@ -168,14 +168,16 @@ bash scripts/uninstall.sh
 
 ```bash
 git clone -b kimi https://github.com/Galaxy-Dawn/claude-scholar.git /tmp/claude-scholar
-mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents
+mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents/subagents
 cp -r /tmp/claude-scholar/skills/research-ideation ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/results-analysis ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/results-report ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/ml-paper-writing ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/review-response ~/.kimi-code/skills/
-cp -r /tmp/claude-scholar/agents/literature-reviewer ~/.kimi-code/agents/
-cp -r /tmp/claude-scholar/agents/paper-miner ~/.kimi-code/agents/
+cp /tmp/claude-scholar/agents/subagents/literature-reviewer.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/literature-reviewer.yaml ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/paper-miner.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/paper-miner.yaml ~/.kimi-code/agents/subagents/
 cp /tmp/claude-scholar/AGENTS.md ~/.kimi-code/AGENTS.md
 cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 ```
@@ -188,8 +190,10 @@ cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 
 ```bash
 git clone -b kimi https://github.com/Galaxy-Dawn/claude-scholar.git /tmp/claude-scholar
+mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents/subagents
 cp -r /tmp/claude-scholar/skills/<skill-name> ~/.kimi-code/skills/
-cp -r /tmp/claude-scholar/agents/<agent-name> ~/.kimi-code/agents/
+cp /tmp/claude-scholar/agents/subagents/<agent-name>.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/<agent-name>.yaml ~/.kimi-code/agents/subagents/
 cp /tmp/claude-scholar/AGENTS.md ~/.kimi-code/AGENTS.md
 cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 ```
@@ -376,6 +380,11 @@ cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 | Skill | `latex-conference-template-organizer` | 乱雑な学会テンプレートをOverleaf対応のライティング構造に整理 |
 | Agent | `paper-miner` | 優れた論文から再利用可能なライティングパターン、構造、学会の期待値を抽出 |
 
+`paper-miner` に論文の書き方を分析させ、インストール済みの共有メモリに保存します。
+関連する知識は `ml-paper-writing`、Nature 系の執筆・推敲、査読対応の Skill
+から利用できます。`scripts/setup.sh` を再実行しても保持され、
+アンインストールでは削除されません。手動で `cp -r` 更新する場合は先にバックアップしてください。
+
 **仕組み**
 - **テンプレート準備**: 学会テンプレートをOverleaf対応構造に整理
 - **ジャーナル寄りの推敲**: 必要に応じて段落ロジック、hedging、section moves を整え、Nature寄りの文体へ近づける

README.md

@@ -378,6 +378,11 @@ Systematic paper writing from template preparation to iterative drafting.
 | Skill | `latex-conference-template-organizer` | Clean messy conference templates into an Overleaf-ready writing structure. |
 | Agent | `paper-miner` | Mine strong papers for reusable writing patterns, structure, and venue expectations. |
 
+Ask `paper-miner` to mine a paper into the installed writing memory. The result
+is shared by `ml-paper-writing`, Nature writing and polishing, and reviewer-response
+skills. Re-running `scripts/setup.sh` preserves the active memory, and uninstall
+leaves it in place. Back up that file before updating skills with a manual `cp -r`.
+
 **How it works**
 - **Template Preparation**: clean conference templates into an Overleaf-ready structure.
 - **Journal-Style Polishing**: tighten paragraph logic, hedging, and section moves for Nature-leaning prose when needed.

README.zh-CN.md

@@ -168,14 +168,16 @@ bash scripts/uninstall.sh
 
 ```bash
 git clone -b kimi https://github.com/Galaxy-Dawn/claude-scholar.git /tmp/claude-scholar
-mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents
+mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents/subagents
 cp -r /tmp/claude-scholar/skills/research-ideation ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/results-analysis ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/results-report ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/ml-paper-writing ~/.kimi-code/skills/
 cp -r /tmp/claude-scholar/skills/review-response ~/.kimi-code/skills/
-cp -r /tmp/claude-scholar/agents/literature-reviewer ~/.kimi-code/agents/
-cp -r /tmp/claude-scholar/agents/paper-miner ~/.kimi-code/agents/
+cp /tmp/claude-scholar/agents/subagents/literature-reviewer.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/literature-reviewer.yaml ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/paper-miner.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/paper-miner.yaml ~/.kimi-code/agents/subagents/
 cp /tmp/claude-scholar/AGENTS.md ~/.kimi-code/AGENTS.md
 cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 ```
@@ -188,8 +190,10 @@ cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 
 ```bash
 git clone -b kimi https://github.com/Galaxy-Dawn/claude-scholar.git /tmp/claude-scholar
+mkdir -p ~/.kimi-code/skills ~/.kimi-code/agents/subagents
 cp -r /tmp/claude-scholar/skills/<skill-name> ~/.kimi-code/skills/
-cp -r /tmp/claude-scholar/agents/<agent-name> ~/.kimi-code/agents/
+cp /tmp/claude-scholar/agents/subagents/<agent-name>.md ~/.kimi-code/agents/subagents/
+cp /tmp/claude-scholar/agents/subagents/<agent-name>.yaml ~/.kimi-code/agents/subagents/
 cp /tmp/claude-scholar/AGENTS.md ~/.kimi-code/AGENTS.md
 cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 ```
@@ -376,6 +380,11 @@ cp /tmp/claude-scholar/AGENTS.zh-CN.md ~/.kimi-code/AGENTS.zh-CN.md
 | Skill | `latex-conference-template-organizer` | 把混乱的会议模板整理成 Overleaf-ready 写作结构。 |
 | Agent | `paper-miner` | 从高质量论文中提炼可复用的写作模式、结构和投稿经验。 |
 
+请让 `paper-miner` 把论文写作模式提炼到安装目录中的共享记忆。
+`ml-paper-writing`、Nature 写作与润色、审稿回复 Skill 都可以读取相关内容。
+重复运行 `scripts/setup.sh` 会保留已写入的记忆，卸载时也不会删除。
+如果手动用 `cp -r` 更新 Skill，请先备份这份记忆。
+
 **工作方式**
 - **模板准备**：把会议模板清理成 Overleaf-ready 结构。
 - **期刊风格润色**：在需要时加强段落逻辑、hedging 和 section moves，使表达更接近 Nature 风格。

agents/subagents/paper-miner.md

@@ -2,6 +2,14 @@
 
 You are a writing knowledge extraction specialist. Your job is to analyze successful papers and extract reusable writing patterns.
 
+Write durable, source-attributed patterns into the active installed
+`~/.kimi-code/skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`.
+Read that file first, merge new insights into its existing sections without
+duplicating sources, and update its Source index. This is the shared memory for
+academic writing skills, not a project-local note. If the installed path differs,
+resolve it from the active Kimi home and report the exact path. Do not write
+into the repository template.
+
 ## Extraction Targets
 
 1. **Structure Patterns**: How is the paper organized? What's the narrative arc?
@@ -30,4 +38,9 @@ For each paper analyzed:
 
 ### For User's Project
 [How these patterns apply to the user's current paper]
-```
+```
+
+After mining, report the exact memory path and sections updated. Preserve
+paper title, venue, year, and source URL or local path. When full text is
+unavailable, mark the extraction as partial and do not infer unsupported
+venue rules.

scripts/setup.sh

@@ -548,6 +548,22 @@ copy_file_safely() {
 
   ensure_parent_dir "$target_file"
 
+  # Mined writing knowledge is user data, not an installer-managed skill file.
+  if [ "$target_file" = "$KIMI_HOME/skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md" ]; then
+    if [ -e "$target_file" ] || [ -L "$target_file" ]; then
+      [ -f "$target_file" ] || [ -L "$target_file" ] || error "Writing memory path is not a file: $target_file"
+      SKIPPED_COUNT=$((SKIPPED_COUNT + 1))
+      return 0
+    fi
+    if [ "$DRY_RUN" = "1" ]; then
+      UPDATED_COUNT=$((UPDATED_COUNT + 1))
+      return 0
+    fi
+    cp -p "$src_file" "$target_file" || error "Failed to initialize writing memory at $target_file"
+    UPDATED_COUNT=$((UPDATED_COUNT + 1))
+    return 0
+  fi
+
   # Unchanged + previously managed → just re-record, skip copy
   if [ -f "$target_file" ] && cmp -s "$src_file" "$target_file"; then
     if was_previously_managed "$target_file"; then

scripts/test_install_uninstall.sh

@@ -46,17 +46,38 @@ test_roundtrip_existing_config() {
   test -f "$home/.kimi/.kimi-scholar-install-state"
   test -f "$home/.kimi/AGENTS.md"
   test -f "$home/.kimi/skills/research-ideation/references/research-contract.md"
+  test -f "$home/.kimi/skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md"
   grep -Fxq "skills/research-ideation/references/research-contract.md" "$home/.kimi/.kimi-scholar-manifest.txt"
 
   run_uninstall "$home"
   test ! -f "$home/.kimi/.kimi-scholar-manifest.txt"
   test ! -f "$home/.kimi/.kimi-scholar-install-state"
   test -f "$home/.kimi/config.toml"
+  test -f "$home/.kimi/skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md"
   ! grep -q '\[agents\.' "$home/.kimi/config.toml"
   ! grep -q '\[mcp_servers\.zotero' "$home/.kimi/config.toml"
   pass "roundtrip with existing config"
 }
 
+test_writing_memory_survives_update_and_legacy_uninstall() {
+  local home memory
+  home="$(make_home)"
+  write_base_config "$home"
+  run_setup "$home"
+  memory="$home/.kimi/skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md"
+  printf '\n### Kept example\n**Source:** Test paper\n' >> "$memory"
+
+  run_setup "$home"
+  grep -Fq '### Kept example' "$memory"
+  ! grep -Fxq 'skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md' "$home/.kimi/.kimi-scholar-manifest.txt"
+
+  # Older manifests may still claim ownership of the file.
+  printf '%s\n' 'skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md' >> "$home/.kimi/.kimi-scholar-manifest.txt"
+  run_uninstall "$home"
+  grep -Fq '### Kept example' "$memory"
+  pass "mined writing memory survives update and legacy uninstall"
+}
+
 test_preserve_existing_mcp_section() {
   local home
   home="$(make_home)"
@@ -642,6 +663,7 @@ main() {
   bash -n "$SETUP_SH"
   bash -n "$UNINSTALL_SH"
   test_roundtrip_existing_config
+  test_writing_memory_survives_update_and_legacy_uninstall
   test_preserve_existing_mcp_section
   test_manifest_missing_skips_safely
   test_identical_preexisting_file_is_not_owned

scripts/uninstall.sh

@@ -380,6 +380,7 @@ collect_targets_to_remove() {
   local rel
   while IFS= read -r rel; do
     [ -n "$rel" ] || continue
+    [ "$rel" = "skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md" ] && continue
     case "$rel" in
       .*|*..*|/*|*\\*|*\$*|*~*) continue ;;
     esac

skills/ml-paper-writing/SKILL.md

@@ -315,7 +315,8 @@ For detailed literature research guidance:
 
 ## Knowledge Base: Paper-Miner Installed Writing Memory
 
-This skill consumes the **active installed writing memory** maintained by `paper-miner`:
+This skill shares the **active installed writing memory** maintained by `paper-miner`
+with the Nature writing, polishing, response, and other academic writing skills:
 
 - `references/knowledge/paper-miner-writing-memory.md`
 
@@ -349,7 +350,7 @@ Extracts reusable writing signals
 ↓
 Updates paper-miner-writing-memory.md
 ↓
-ml-paper-writing reuses that memory later
+Relevant academic writing skills reuse that memory later
 ```
 
 ### When to use this memory
@@ -372,8 +373,8 @@ When drafting or revising with `ml-paper-writing`, read this memory **before** w
 - venue-facing polishing.
 
 Use this read order:
-1. `references/knowledge/paper-miner-writing-memory.md`
-2. repo-local evidence and experiment artifacts
+1. repo-local evidence and experiment artifacts
+2. `references/knowledge/paper-miner-writing-memory.md`, when relevant entries exist
 3. cited papers or notes if needed
 4. venue template and formatting constraints
 

skills/ml-paper-writing/references/knowledge/README.md

@@ -8,7 +8,8 @@ The canonical paper-miner memory is:
 
 - `paper-miner-writing-memory.md`
 
-This is the **only maintained paper-miner writing memory**.
+This is the **only maintained paper-miner writing memory**. All academic writing
+skills in the same client installation may read relevant entries from it.
 
 It stores:
 - writing patterns mined,
@@ -27,6 +28,12 @@ This memory is:
 - **cross-project within this Kimi install**,
 - **not project-specific**.
 
+The installer must preserve this file across updates, and the uninstaller must
+leave it in place. It contains user-mined data, even though the repository also
+ships a starter template at the same relative path.
+When updating by hand with `cp -r`, back up this file first and restore it
+after the copy. The installer's preservation rule does not apply to manual copies.
+
 If `paper-miner` is invoked inside a project, it may use project context to understand relevance, but it still writes only to the active installed skill memory.
 
 ## Legacy files

skills/nature-data/SKILL.md

@@ -18,6 +18,13 @@ flags.
 The governing policy layer is Springer Nature / Nature Portfolio data policy. The implementation
 layer is FAIR data practice and DataCite-style citation metadata.
 
+For academic wording only, the active installed
+`skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`
+under the current client's skill home may offer relevant phrasing examples.
+Use it only when it contains source-attributed availability language. Journal
+policy and the author's actual data-access facts take precedence. If the memory
+is absent or irrelevant, continue without it.
+
 ## Chinese-user operating mode
 
 When the user writes in Chinese, provides a Chinese manuscript note, or asks for "中文对应",

skills/nature-polishing/SKILL.md

@@ -26,6 +26,16 @@ the curated article-pattern reference. The wording layer should come from
 - If the draft is Chinese or structurally rough, reconstruct the logic first and the prose second.
 - Avoid em dashes in polished output by default. Prefer commas, parentheses, or full stops. Use colons sparingly unless the user explicitly asks to preserve dash-based punctuation or wants a colon-led style.
 
+## Mined writing memory
+
+For academic prose, check the active installed
+`skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`
+under the current client's skill home. Use only source-attributed entries that
+fit the section and venue, alongside this skill's curated references. Treat them
+as optional style and structure examples. Preserve the author's claims and
+evidence, follow current journal instructions, and do not copy source phrasing.
+If the memory has no relevant entries, continue without it.
+
 ## When to open extra files
 
 These files are reference support. Use them after the section's rhetorical job is clear.

skills/nature-response/SKILL.md

@@ -30,6 +30,15 @@ justified scientific response, or unresolved author action.
 - When a reviewer misunderstood the manuscript, first consider whether the manuscript presentation caused the misunderstanding.
 - Treat rebuttal letters as potentially public review artifacts; write with professional tone and traceability.
 
+## Mined writing memory
+
+For response structure or tone, check the active installed
+`skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`
+under the current client's skill home. Read only relevant rebuttal and venue
+entries. Reviewer comments, manuscript changes, and journal instructions remain
+the source of truth. Do not borrow claims or copy source phrasing. If the memory
+is absent or has no relevant entries, continue with this skill's references.
+
 ## Accepted inputs
 
 The skill may receive:

skills/nature-writing/SKILL.md

@@ -20,6 +20,16 @@ not merely polishing existing sentences.
 - If essential evidence is missing, write a placeholder or ask for the missing
   input instead of filling the gap.
 
+## Mined writing memory
+
+Before drafting or restructuring academic prose, check the active installed
+`skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`
+under the current client's skill home. Read only entries relevant to this paper's
+section, article type, and venue. Use mined patterns for structure and wording
+ideas, while grounding every claim in the author's evidence and the target
+journal's requirements. Do not copy source phrasing. If the memory is absent or
+has no relevant entries, continue with this skill's references.
+
 ## When to open extra files
 
 | File | Open when |

skills/writing-anti-ai/SKILL.md

@@ -11,6 +11,13 @@ tags: [Writing, AI, Anti-AI, Humanizer]
 
 Remove AI-generated writing patterns from text to make it sound natural and human-written. Supports both English and Chinese.
 
+When editing academic prose, check the active installed
+`skills/ml-paper-writing/references/knowledge/paper-miner-writing-memory.md`
+under the current client's skill home for relevant source-attributed writing
+examples. Use it only to guide natural structure and rhythm. Keep the author's
+meaning, evidence, and target venue constraints intact. If the memory is absent
+or irrelevant, continue with this skill's own pattern references.
+
 ## Overview
 
 This skill identifies and eliminates predictable AI writing patterns from prose, based on [Wikipedia: Signs of AI writing](https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing), maintained by WikiProject AI Cleanup.