@fitlab-ai/agent-infra 0.11.4 → 0.11.6

package/templates/.agents/skills/analyze-task/SKILL.en.md

@@ -15,7 +15,7 @@ If the entry operands contain `--orchestrated`, bind `{execution-flag}` to `--or
 
 ### Flow decision
 
-The analysis artifact must record the selected path, factual basis, unmet higher-path conditions, and factual upgrade triggers in `## Flow Decision`. The path is streamlined, standard, or full; file count, module count, and speculative risk cannot upgrade it by themselves.
+The analysis artifact must record `Path`, `Basis`, `Unmet Higher-path Conditions`, and `Upgrade Triggers` exactly once each, with non-empty values, in `## Flow Decision`. The path is streamlined, standard, or full; file count, module count, and speculative risk cannot upgrade it by themselves. The analysis `finalize-local` reuses the lifecycle-path parser to validate this decision; directly repair the formal artifact and rerun it for missing, duplicate, or invalid values.
 
 ### Persisted Report Evidence
 
@@ -160,7 +160,7 @@ Create `.agents/workspace/active/{task-id}/{analysis-artifact}`.
 
 ## State Check
 
-> Record the Step 0 state-check command, task/artifact scope, key result, and uncovered parts; each command starts with `$ `. Do not paste complete directory listings or task tails on normal success; retain decisive raw lines only for failures, blocking conditions, identity mismatches, or disputes.
+> Record the Step 0 state-check command, task/artifact scope, key result, and uncovered parts. Do not paste complete directory listings or task tails on normal success; retain decisive raw lines only for failures, blocking conditions, identity mismatches, or disputes.
 
 ## Requirement Source
 
@@ -187,6 +187,15 @@ Create `.agents/workspace/active/{task-id}/{analysis-artifact}`.
 ## Dependencies
 - {Required dependencies and coordination with other modules}
 
+<!-- lifecycle-path-decision-template:start -->
+## Flow Decision
+
+- **Path**: {lifecycle-path}
+- **Basis**: {Factual basis for selecting the current path}
+- **Unmet Higher-path Conditions**: {Factual conditions for the higher path that are not met; for full, state that no higher path exists}
+- **Upgrade Triggers**: {Factual triggers that require reassessing the path}
+<!-- lifecycle-path-decision-template:end -->
+
 ## Assumptions
 
 > If this analysis depends on assumptions, list them here; omit this section if there are none.

package/templates/.agents/skills/analyze-task/SKILL.zh-CN.md

@@ -15,7 +15,7 @@ description: >
 
 ### 流程裁定
 
-分析产物必须在 `## 流程裁定` 中记录本任务路径、判定依据、未满足的更高路径条件和事实型升级触发条件。路径仅可为精简、标准或完整；文件数、模块数和推测性风险不能单独升级路径。
+分析产物必须在 `## 流程裁定` 中记录 `本任务路径`、`判定依据`、`未满足的更高路径条件` 和 `升级触发条件` 四个字段，每个字段恰好一次且非空。路径仅可为精简、标准或完整；文件数、模块数和推测性风险不能单独升级路径。analysis 的 `finalize-local` 会复用生命周期路径解析器校验该决定；缺失、重复或无效值必须直接修复正式产物后重跑。
 
 ### 持久化报告证据
 
@@ -163,7 +163,7 @@ agent-infra-internal task-artifact {task-id} init --family analysis --artifact {
 
 ## 状态核对
 
-> 记录第 0 步状态核对命令、任务/产物范围、关键结果和未覆盖部分；每条命令以 `$ ` 开头。正常成功不粘贴完整目录清单或 `task.md` 尾部，失败、阻塞、身份不一致或争议时附决定性原文行。
+> 记录第 0 步状态核对命令、任务/产物范围、关键结果和未覆盖部分。正常成功不粘贴完整目录清单或 `task.md` 尾部，失败、阻塞、身份不一致或争议时附决定性原文行。
 
 ## 需求来源
 
@@ -190,6 +190,15 @@ agent-infra-internal task-artifact {task-id} init --family analysis --artifact {
 ## 依赖关系
 - {需要的依赖和与其他模块的协调}
 
+<!-- lifecycle-path-decision-template:start -->
+## 流程裁定
+
+- **本任务路径**：{lifecycle-path}
+- **判定依据**：{选择当前路径的事实依据}
+- **未满足的更高路径条件**：{当前未满足的更高路径条件；完整路径写不存在更高路径的事实}
+- **升级触发条件**：{未来需要重新评估路径的事实触发条件}
+<!-- lifecycle-path-decision-template:end -->
+
 ## 资格审计
 
 > 按 `.agents/rules/decision-qualification.md` 填写约束依赖、候选资格、分类结果、上游关系、依赖快照五张表；没有 artifact 上游时保留空表头，不虚构关系。

package/templates/.agents/skills/archive-tasks/SKILL.en.md

@@ -7,7 +7,7 @@ description: >
 
 # Archive Completed Tasks
 
-Move completed tasks from `.agents/workspace/completed/` into `.agents/workspace/archive/YYYY/MM/DD/TASK-xxx/` and rebuild a three-level archive index:
+Archive completed tasks from `.agents/workspace/completed/` into `.agents/workspace/archive/YYYY/MM/DD/TASK-xxx/local/` and rebuild a three-level archive index. Local task materials live in `local/` with a path-sorted `contents.sha256`; the TASK root contains no ordinary files.
 - root manifest: `.agents/workspace/archive/manifest.md`
 - yearly manifest: `.agents/workspace/archive/YYYY/manifest.md`
 - monthly manifest: `.agents/workspace/archive/YYYY/MM/manifest.md`
@@ -32,7 +32,7 @@ bash .agents/skills/archive-tasks/scripts/archive-tasks.sh [--days N | --before
 
 The script is responsible for:
 - reading `completed_at` from `task.md` frontmatter and falling back to `updated_at`
-- moving task directories directly into `YYYY/MM/DD/TASK-xxx/` without compression
+- storing local task materials under `YYYY/MM/DD/TASK-xxx/local/` without compression
 - skipping already archived, missing, or malformed tasks
 - rebuilding root, yearly, and monthly manifests from all archived tasks
 - printing an archive and skip summary

package/templates/.agents/skills/archive-tasks/SKILL.zh-CN.md

@@ -7,7 +7,7 @@ description: >
 
 # 归档已完成任务
 
-将 `.agents/workspace/completed/` 中的已完成任务移动到 `.agents/workspace/archive/YYYY/MM/DD/TASK-xxx/`，并重建三级归档索引：
+将 `.agents/workspace/completed/` 中的已完成任务归档到 `.agents/workspace/archive/YYYY/MM/DD/TASK-xxx/local/`，并重建三级归档索引。TASK 本地材料位于 `local/`，含按路径排序的 `contents.sha256`；TASK 根目录不存放普通文件。
 - 根 manifest：`.agents/workspace/archive/manifest.md`
 - 年 manifest：`.agents/workspace/archive/YYYY/manifest.md`
 - 月 manifest：`.agents/workspace/archive/YYYY/MM/manifest.md`
@@ -32,7 +32,7 @@ bash .agents/skills/archive-tasks/scripts/archive-tasks.sh [--days N | --before
 
 脚本负责：
 - 解析 `task.md` frontmatter 中的 `completed_at`（缺失时回退到 `updated_at`）
-- 按 `YYYY/MM/DD/TASK-xxx/` 目录直接移动任务，不压缩
+- 按 `YYYY/MM/DD/TASK-xxx/local/` 保存任务本地材料，不压缩
 - 跳过已归档、缺少元数据或不存在的任务
 - 全量重建根 / 年 / 月三级 manifest
 - 输出归档与跳过摘要

package/templates/.agents/skills/archive-tasks/scripts/archive-tasks.sh

@@ -8,9 +8,14 @@ WORKSPACE_ROOT="$REPO_ROOT/.agents/workspace"
 COMPLETED_DIR="$WORKSPACE_ROOT/completed"
 ARCHIVE_DIR="$WORKSPACE_ROOT/archive"
 MANIFEST_PATH="$ARCHIVE_DIR/manifest.md"
+tmpdir=""
+cleanup() {
+  [ -z "$tmpdir" ] || rm -rf "$tmpdir"
+}
+trap 'cleanup' 0
+trap 'cleanup; exit 1' HUP INT TERM
 
-tmpdir="$(mktemp -d)"
-trap 'rm -rf "$tmpdir"' EXIT HUP INT TERM
+tmpdir="$(mktemp -d "$WORKSPACE_ROOT/.archive-tmp.XXXXXX")"
 
 IDS_FILE="$tmpdir/task-ids.txt"
 : > "$IDS_FILE"
@@ -160,6 +165,17 @@ archive_task_dir() {
     return 0
   fi
 
+  find "$task_dir" -name manifest.md -print > "$tmpdir/task-manifests.txt"
+  if [ -s "$tmpdir/task-manifests.txt" ]; then
+    echo "TASK content cannot contain manifest.md: $task_id" >&2
+    return 1
+  fi
+  find "$task_dir" ! -type f ! -type d -print > "$tmpdir/task-special-files.txt"
+  if [ -s "$tmpdir/task-special-files.txt" ]; then
+    echo "TASK content cannot contain symbolic links or special files: $task_id" >&2
+    return 1
+  fi
+
   completed_at=$(extract_completed_at "$task_file")
   task_date=$(printf '%s' "$completed_at" | cut -c1-10)
 
@@ -179,12 +195,89 @@ archive_task_dir() {
     return 0
   fi
 
-  mkdir -p "$ARCHIVE_DIR/$year/$month/$day"
-  mv "$task_dir" "$destination_dir"
+  mkdir -p "$(dirname "$destination_dir")"
+  staged_dir="$tmpdir/$task_id"
+  mkdir -p "$staged_dir/local"
+  cp -R "$task_dir"/. "$staged_dir/local/"
+  write_contents_hash "$staged_dir/local"
+  mv "$staged_dir" "$destination_dir"
+  if ! rm -rf "$task_dir"; then
+    if ! rm -rf "$destination_dir"; then
+      echo "Failed to remove completed task and roll back archive copy: $task_id" >&2
+    else
+      echo "Failed to remove completed task; rolled back archive copy: $task_id" >&2
+    fi
+    return 1
+  fi
   archived_count=$((archived_count + 1))
   printf 'Archived %s -> %s\n' "$task_id" "$relative_path"
 }
 
+write_contents_hash() {
+  local_dir="$1"
+  hash_path=${2:-$local_dir/contents.sha256}
+  hash_tmp="$tmpdir/contents.sha256"
+  newline='
+'
+  : > "$hash_tmp"
+  find "$local_dir" -type f ! -name contents.sha256 | LC_ALL=C sort | while IFS= read -r source_file; do
+    relative_file=${source_file#"$local_dir"/}
+    case "$relative_file" in *"$newline"*|*"$(printf '\r')"*) echo "Newline path is not supported" >&2; return 1 ;; esac
+    if command -v sha256sum >/dev/null 2>&1; then
+      digest=$(sha256sum "$source_file" | awk '{print $1}')
+    else
+      digest=$(shasum -a 256 "$source_file" | awk '{print $1}')
+    fi
+    printf '%s  %s\n' "$digest" "$relative_file" >> "$hash_tmp"
+  done
+  mv "$hash_tmp" "$hash_path"
+  if [ "$hash_path" = "$local_dir/contents.sha256" ] && command -v sha256sum >/dev/null 2>&1; then
+    (cd "$local_dir" && sha256sum -c contents.sha256 >/dev/null)
+  elif [ "$hash_path" = "$local_dir/contents.sha256" ]; then
+    (cd "$local_dir" && shasum -a 256 -c contents.sha256 >/dev/null)
+  fi
+}
+
+validate_source_hash() {
+  source_dir="$1"
+  [ -f "$source_dir/task.md" ] || { echo "Archive source lacks task.md: $source_dir" >&2; return 1; }
+  [ -f "$source_dir/contents.sha256" ] || { echo "Archive source lacks contents.sha256: $source_dir" >&2; return 1; }
+  write_contents_hash "$source_dir" "$tmpdir/contents.expected"
+  if ! cmp -s "$tmpdir/contents.expected" "$source_dir/contents.sha256"; then
+    echo "Archive source checksum mismatch: $source_dir" >&2
+    return 1
+  fi
+}
+
+validate_existing_archive() {
+  [ -d "$ARCHIVE_DIR" ] || return 0
+  for year_dir in "$ARCHIVE_DIR"/[0-9][0-9][0-9][0-9]; do
+    [ -d "$year_dir" ] || continue
+    for month_dir in "$year_dir"/[0-9][0-9]; do
+      [ -d "$month_dir" ] || continue
+      for day_dir in "$month_dir"/[0-9][0-9]; do
+        [ -d "$day_dir" ] || continue
+        for task_dir in "$day_dir"/TASK-*; do
+          [ -d "$task_dir" ] || continue
+          [ -d "$task_dir/local" ] || { echo "Unsupported legacy archive task layout: $task_dir" >&2; return 1; }
+          for task_entry in "$task_dir"/*; do
+            [ -e "$task_entry" ] || continue
+            case "$(basename "$task_entry")" in
+              local|github|derived) [ -d "$task_entry" ] || { echo "Invalid archive source directory: $task_entry" >&2; return 1; } ;;
+              *) echo "Unexpected TASK root entry: $task_entry" >&2; return 1 ;;
+            esac
+          done
+          find "$task_dir/local" -name manifest.md -print > "$tmpdir/task-manifests.txt"
+          [ ! -s "$tmpdir/task-manifests.txt" ] || { echo "TASK content cannot contain manifest.md: $task_dir" >&2; return 1; }
+          find "$task_dir/local" ! -type f ! -type d -print > "$tmpdir/task-special-files.txt"
+          [ ! -s "$tmpdir/task-special-files.txt" ] || { echo "TASK content cannot contain symbolic links or special files: $task_dir" >&2; return 1; }
+          validate_source_hash "$task_dir/local"
+        done
+      done
+    done
+  done
+}
+
 should_archive_filtered_task() {
   task_dir="$1"
   task_file="$task_dir/task.md"
@@ -231,7 +324,7 @@ rebuild_manifest() {
           [ -d "$task_dir" ] || continue
 
           task_id=$(basename "$task_dir")
-          task_file="$task_dir/task.md"
+          task_file="$task_dir/local/task.md"
           relative_path="$year/$month/$day/$task_id/"
           title="$task_id"
           task_type="unknown"
@@ -261,7 +354,15 @@ rebuild_manifest() {
     done
   done
 
-  find "$ARCHIVE_DIR" -type f -name 'manifest.md' -exec rm -f {} \;
+  rm -f "$ARCHIVE_DIR/manifest.md"
+  for year_dir in "$ARCHIVE_DIR"/[0-9][0-9][0-9][0-9]; do
+    [ -d "$year_dir" ] || continue
+    rm -f "$year_dir/manifest.md"
+    for month_dir in "$year_dir"/[0-9][0-9]; do
+      [ -d "$month_dir" ] || continue
+      rm -f "$month_dir/manifest.md"
+    done
+  done
 
   awk -F'\t' '{print $1 "\t" $2}' "$entries_file" | LC_ALL=C sort -u > "$month_keys_file"
   awk -F'\t' '{print $1}' "$entries_file" | LC_ALL=C sort -u > "$year_keys_file"
@@ -436,6 +537,8 @@ if [ -s "$IDS_FILE" ]; then
   MODE="ids"
 fi
 
+validate_existing_archive
+
 archived_count=0
 skipped_count=0
 

package/templates/.agents/skills/code-task/reference/report-template.en.md

@@ -102,25 +102,25 @@ Use this structure when creating `code.md` or `code-r{N}.md`.
 
 ## State Check
 
-> Record the state-check command, task/artifact scope, key result, and uncovered parts; each command starts with `$ `.
+> Record the state-check command, task/artifact scope, key result, and uncovered parts.
 > Follow `.agents/rules/evidence-reporting.md`: record task/artifact scope, key result, and uncovered parts; do not paste complete successful stdout.
 
-Record at least one actually executed `$ ` command and its result summary; prose-only status is insufficient. An implementation-stage report normally includes:
+Record at least one actually executed command and its result summary; prose-only status is insufficient. An implementation-stage report normally includes:
 
 ```text
-$ agent-infra-internal task-snapshot {task-id} --format text
+agent-infra-internal task-snapshot {task-id} --format text
 {task state, implementation-artifact scope, and uncovered-area summary}
-$ agent-infra-internal task-artifact {task-id} inspect --family code
+agent-infra-internal task-artifact {task-id} inspect --family code
 {code artifact, upstream input, and next-step summary}
 ```
 
 ## Raw Evidence
 
-> Follow `.agents/rules/evidence-reporting.md`: pair each claim with a `$ ` command and proportionate result summary; include a decisive raw excerpt only for failures, blocking conditions, or disputes.
+> Follow `.agents/rules/evidence-reporting.md`: pair each claim with an executable command and proportionate result summary; include a decisive raw excerpt only for failures, blocking conditions, or disputes.
 
 - Claim: {verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```
 

package/templates/.agents/skills/code-task/reference/report-template.zh-CN.md

@@ -102,25 +102,25 @@
 
 ## 状态核对
 
-> 记录状态核对命令、任务/产物范围、关键结果和未覆盖部分；每条命令以 `$ ` 开头。
+> 记录状态核对命令、任务/产物范围、关键结果和未覆盖部分。
 > 按 `.agents/rules/evidence-reporting.md` 同时记录任务/产物范围、关键结果和未覆盖部分；正常成功不粘贴完整 stdout。
 
-必须记录至少一条实际执行的 `$ ` 命令及其结果摘要，不能只写自然语言状态。实现阶段通常包括：
+必须记录至少一条实际执行的命令及其结果摘要，不能只写自然语言状态。实现阶段通常包括：
 
 ```text
-$ agent-infra-internal task-snapshot {task-id} --format text
+agent-infra-internal task-snapshot {task-id} --format text
 {任务状态、实现产物范围和未覆盖项摘要}
-$ agent-infra-internal task-artifact {task-id} inspect --family code
+agent-infra-internal task-artifact {task-id} inspect --family code
 {code artifact、上游输入和下一步摘要}
 ```
 
 ## 证据原文
 
-> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对 `$ ` 命令和相称结果摘要；仅为失败、阻塞或争议附决定性原文摘录。
+> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对可执行命令和相称结果摘要；仅为失败、阻塞或争议附决定性原文摘录。
 
 - 断言：{verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```
 

package/templates/.agents/skills/complete-manual-validation/config/verify.en.json

@@ -23,8 +23,8 @@
         "Validation Details",
         "PR Summary Sync"
       ],
-      "required_patterns": [
-        "^\\$ "
+      "state_check_patterns": [
+        "agent-infra-internal[ \\t]+task-snapshot\\b"
       ]
     },
     "activity-log": {

package/templates/.agents/skills/complete-manual-validation/config/verify.zh-CN.json

@@ -23,8 +23,8 @@
         "验证详情",
         "PR 摘要同步"
       ],
-      "required_patterns": [
-        "^\\$ "
+      "state_check_patterns": [
+        "agent-infra-internal[ \\t]+task-snapshot\\b"
       ]
     },
     "activity-log": {

package/templates/.agents/skills/complete-manual-validation/reference/report-template.en.md

@@ -12,9 +12,7 @@ Read this file before creating `manual-validation.md` / `manual-validation-r{N}.
 ## State Check
 
 ```text
-$ git status -s
-$ ls -la .agents/workspace/active/{task-id}/
-$ tail .agents/workspace/active/{task-id}/task.md
+agent-infra-internal task-snapshot {task-id} --format text
 ```
 > Follow `.agents/rules/evidence-reporting.md`: record task scope, key result, and uncovered parts; do not paste the complete directory listing or task tail for normal success.
 

package/templates/.agents/skills/complete-manual-validation/reference/report-template.zh-CN.md

@@ -12,9 +12,7 @@
 ## 状态核对
 
 ```text
-$ git status -s
-$ ls -la .agents/workspace/active/{task-id}/
-$ tail .agents/workspace/active/{task-id}/task.md
+agent-infra-internal task-snapshot {task-id} --format text
 ```
 > 按 `.agents/rules/evidence-reporting.md` 记录任务范围、关键结果和未覆盖部分；正常成功不粘贴完整目录清单或 task tail。
 

package/templates/.agents/skills/complete-task/config/verify.en.json

@@ -38,8 +38,8 @@
       "required_sections": [
         "State Check"
       ],
-      "required_patterns": [
-        "^\\$ "
+      "state_check_patterns": [
+        "agent-infra-internal[ \\t]+task-snapshot\\b"
       ]
     }
   }

package/templates/.agents/skills/complete-task/config/verify.zh-CN.json

@@ -38,8 +38,8 @@
       "required_sections": [
         "状态核对"
       ],
-      "required_patterns": [
-        "^\\$ "
+      "state_check_patterns": [
+        "agent-infra-internal[ \\t]+task-snapshot\\b"
       ]
     }
   }

package/templates/.agents/skills/create-release-note/SKILL.en.md

@@ -53,6 +53,7 @@ agent-infra-internal platform-release-notes context \
 
 **Purpose**:
 - Part A: Analyze the section structure, heading style, emoji usage, and item format from the latest 3 historical release notes
+- Each `history` item contains a tag, body, and Release URL; use the body as a format sample, not as current change facts
 - Part B: Provide a static complete category list so no existing category is omitted
 - This static list ensures existing category names are not missed during classification; if the current release has no entries for a category, Step 7 still omits the empty section
 - When generating release notes in Step 7, **must** follow both the historical format style and the full category list gathered in Step 3
@@ -60,7 +61,7 @@ agent-infra-internal platform-release-notes context \
 
 ### 4. Collect Merged PRs and Contributors
 
-Use `pullRequests` and `commits` from Step 3. Each commit's `authors` are normalized platform facts containing the git author and co-authors; the skill does not interpret raw platform fields or email rules.
+Use `pullRequests` and `commits` from Step 3. Each commit's `authors` contains only that commit's platform authors; `pullRequestNumbers` records the platform's PR associations. The skill does not interpret raw platform fields or email rules. A PR's `closingIssues` contains its actual closing Issues; never infer a relationship from matching PR and Issue numbers.
 
 ### 5. Collect Related Issues
 
@@ -79,31 +80,36 @@ Use each PR's `closingIssues` from Step 3. Do not parse platform-specific refere
 
 ### 7. Generate Release Notes
 
-**Prioritize the historical format style obtained in Step 3 and ensure all categories listed in Step 3 are covered.** If historical release notes exist, strictly follow their section structure, heading style (including emojis), item format, and bilingual layout.
+**Prioritize the historical format style obtained in Step 3 and ensure all categories listed in Step 3 are covered.** Dynamically use the latest 3 published notes each time. Treat sections, bilingual layout, and installation instructions shared by those notes as stable format; allow item content to vary by release. Never copy an example's authors or links into the current release.
 
 If no historical release notes exist, use the following default Markdown format:
 
 ```markdown
-## {Module/Platform Name}
+## {Module / 模块}
 
-### Enhancement
+### Enhancement / 增强
 
 - [{scope}] Description by @author in [#N](url)
 
-### Bugfix
+### Bugfix / 修复
 
 - [{scope}] Description by @author in [#N](url)
 
-## Contributors
+## Contributors / 贡献者
 
 @contributor1, @contributor2, @contributor3, @reporter1 (reported #N)
+
+## Installation / 安装
+
+{English and Chinese installation instructions for this release, following stable steps in historical notes}
 ```
 
 **Format rules**:
-1. Item format: `- [scope] Description by @author in [#N](url)`
-2. Issue + PR: `in [#Issue](url) and [#PR](url)`
-3. Description: Use PR title, remove `type(scope):` prefix, capitalize first letter
-4. **Contributor collection**:
+1. PR item format: `- [scope] Description by @author in [#N](url)`
+2. Issue + PR: `in [#Issue](url) and [#PR](url)`; Issue reporters come from that PR's `closingIssues`
+3. Render each commit with an empty `pullRequestNumbers` as a standalone item using its subject and resolvable author; link to the commit itself when its URL is available, without adding an Issue or PR link
+4. Description: Use the PR title or commit subject; remove `type(scope):` from PR titles and capitalize the first letter
+5. **Contributor collection**:
    - **Data sources**:
      - PR authors from the typed context
      - Commit co-authors from the typed context's commit `authors`
@@ -123,7 +129,7 @@ If no historical release notes exist, use the following default Markdown format:
      - Reporter-only contributors use the format `@login (reported #N)`; if the same reporter reported multiple Issues, use `@login (reported #N1, #N2)`
      - Reporters are appended after code contributors in the Contributors section, separated by commas
      - Sort reporters by reported Issue count descending, then lexicographically by login for ties
-5. Empty sections: Omit sections with no entries
+6. Empty sections: Omit categories with no entries; retain bilingual headings and the installation section from the historical format
 
 ### 8. Stage, Present, and Confirm
 

package/templates/.agents/skills/create-release-note/SKILL.zh-CN.md

@@ -53,6 +53,7 @@ agent-infra-internal platform-release-notes context \
 
 **用途**：
 - Part A：分析最近 3 条历史发布说明的章节结构、标题风格、emoji 使用、条目格式
+- 每项 `history` 含 tag、正文和 Release URL；正文只作为格式样例，不作为当前版本的变更事实
 - Part B：提供静态完整分类清单，确保后续生成时不遗漏已有分类
 - 该静态清单用于确保变更分类时不遗漏已有类别名称；若当前版本无该类变更，仍按步骤 7 的格式规则省略空分类
 - 后续步骤 7 生成发布说明时，**必须**同时参考步骤 3 的历史格式风格和完整分类清单，保持版本间的一致性
@@ -60,7 +61,7 @@ agent-infra-internal platform-release-notes context \
 
 ### 4. 收集已合并的 PR 与贡献者
 
-使用步骤 3 返回的 `pullRequests` 与 `commits`。每个 commit 的 `authors` 已按平台事实规范化并包含 git author 与 co-author；技能不读取平台原始字段或邮箱规则。
+使用步骤 3 返回的 `pullRequests` 与 `commits`。每个 commit 的 `authors` 只包含该 commit 的平台作者；`pullRequestNumbers` 表示平台记录的 PR 关联。技能不读取平台原始字段或邮箱规则。PR 的 `closingIssues` 是该 PR 实际关闭的 Issue，不按 PR/Issue 编号推断关联。
 
 ### 5. 收集关联 Issue
 
@@ -79,31 +80,36 @@ agent-infra-internal platform-release-notes context \
 
 ### 7. 生成发布说明
 
-**优先使用步骤 3 中获取的历史格式风格，并确保覆盖步骤 3 列出的所有分类。** 如果存在历史发布说明，严格沿用其章节结构、标题风格（含 emoji）、条目格式和双语布局。
+**优先使用步骤 3 中获取的历史格式风格，并确保覆盖步骤 3 列出的所有分类。** 每次动态参考最近 3 条已发布说明；把这些正文共同使用的章节、双语布局和安装段作为稳定格式，条目内容按当前版本变化。不要把样例中的作者或链接复制到当前版本。
 
 如果没有历史发布说明，使用以下默认格式化为 Markdown：
 
 ```markdown
-## {模块/平台名称}
+## {Module / 模块}
 
-### Enhancement
+### Enhancement / 增强
 
 - [{scope}] Description by @author in [#N](url)
 
-### Bugfix
+### Bugfix / 修复
 
 - [{scope}] Description by @author in [#N](url)
 
-## Contributors
+## Contributors / 贡献者
 
 @contributor1, @contributor2, @contributor3, @reporter1 (reported #N)
+
+## Installation / 安装
+
+{本版本的中英文安装说明，沿用历史正文中的稳定安装步骤}
 ```
 
 **格式规则**：
-1. 条目格式：`- [scope] Description by @author in [#N](url)`
-2. Issue + PR：`in [#Issue](url) and [#PR](url)`
-3. 描述：使用 PR 标题，移除 `type(scope):` 前缀，首字母大写
-4. **贡献者搜集**：
+1. PR 条目格式：`- [scope] Description by @author in [#N](url)`
+2. Issue + PR：`in [#Issue](url) and [#PR](url)`；Issue reporter 来自该 PR 的 `closingIssues`
+3. `pullRequestNumbers` 为空的 commit 单独作为条目，使用 commit subject 和可解析作者；有 commit URL 时链接 commit 本身，不添加 Issue/PR 链接
+4. 描述：使用 PR 标题或 commit subject；移除 PR 标题中的 `type(scope):` 前缀，首字母大写
+5. **贡献者搜集**：
    - **数据源**：
      - PR author：来自 `.agents/rules/release-commands.md` 中已合并 PR 查询规则
      - Commit co-authors：来自步骤 3 typed context 的 commit `authors`
@@ -123,7 +129,7 @@ agent-infra-internal platform-release-notes context \
      - 仅报告贡献的用户以 `@login (reported #N)` 格式展示；同一 reporter 报告多个 Issue 时使用 `@login (reported #N1, #N2)`
      - Reporter 在 Contributors 段落中排在代码贡献者之后，以逗号分隔追加
      - Reporter 之间按报告的 Issue 数量降序排列，数量相同时按 login 字典序
-5. 空部分：省略没有条目的部分
+6. 空部分：省略没有条目的分类；保留历史格式中的双语章节与安装段
 
 ### 8. Stage、展示并确认
 

package/templates/.agents/skills/review-analysis/reference/report-template.en.md

@@ -147,16 +147,16 @@ Use this template when writing `review-analysis.md` or `review-analysis-r{N}.md`
 
 ## State Check
 
-> Record the state-check command, review scope, key result, and uncovered parts; each command starts with `$ `.
+> Record the state-check command, review scope, key result, and uncovered parts.
 > Follow `.agents/rules/evidence-reporting.md`: record review scope, key result, and uncovered parts; do not paste complete successful stdout.
 
 ## Raw Evidence
 
-> Follow `.agents/rules/evidence-reporting.md`: pair each claim with a `$ ` command and proportionate result summary; Blockers, failures, blocking conditions, or disputes require a reproducible command, exact location, and decisive excerpt; a judgment that cannot be reproduced must be downgraded or moved to Self-Doubt.
+> Follow `.agents/rules/evidence-reporting.md`: pair each claim with an executable command and proportionate result summary; Blockers, failures, blocking conditions, or disputes require a reproducible command, exact location, and decisive excerpt; a judgment that cannot be reproduced must be downgraded or moved to Self-Doubt.
 
 - Claim: {verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```
 

package/templates/.agents/skills/review-analysis/reference/report-template.zh-CN.md

@@ -147,16 +147,16 @@
 
 ## 状态核对
 
-> 记录状态核对命令、审查范围、关键结果和未覆盖部分；每条命令以 `$ ` 开头。
+> 记录状态核对命令、审查范围、关键结果和未覆盖部分。
 > 按 `.agents/rules/evidence-reporting.md` 同时记录审查范围、关键结果和未覆盖部分；正常成功不粘贴完整 stdout。
 
 ## 证据原文
 
-> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对 `$ ` 命令和相称结果摘要；Blocker、失败、阻塞或争议必须保留可复现命令、准确位置和决定性摘录；无法复现的判断须降级或移入「自我质疑」。
+> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对可执行命令和相称结果摘要；Blocker、失败、阻塞或争议必须保留可复现命令、准确位置和决定性摘录；无法复现的判断须降级或移入「自我质疑」。
 
 - 断言：{verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```
 

package/templates/.agents/skills/review-code/reference/report-template.en.md

@@ -153,16 +153,16 @@ Use this template when writing `review-code.md` or `review-code-r{N}.md`.
 
 ## State Check
 
-> Record the state-check command, review scope, key result, and uncovered parts; each command starts with `$ `.
+> Record the state-check command, review scope, key result, and uncovered parts.
 > Follow `.agents/rules/evidence-reporting.md`: record review scope, key result, and uncovered parts; do not paste complete successful stdout.
 
 ## Raw Evidence
 
-> Follow `.agents/rules/evidence-reporting.md`: pair each claim with a `$ ` command and proportionate result summary; Blockers, failures, blocking conditions, or disputes require reproducible evidence, exact location, and a decisive excerpt; a judgment that cannot be reproduced must be downgraded or moved to Self-Doubt.
+> Follow `.agents/rules/evidence-reporting.md`: pair each claim with an executable command and proportionate result summary; Blockers, failures, blocking conditions, or disputes require reproducible evidence, exact location, and a decisive excerpt; a judgment that cannot be reproduced must be downgraded or moved to Self-Doubt.
 
 - Claim: {verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```
 

package/templates/.agents/skills/review-code/reference/report-template.zh-CN.md

@@ -153,16 +153,16 @@
 
 ## 状态核对
 
-> 记录状态核对命令、审查范围、关键结果和未覆盖部分；每条命令以 `$ ` 开头。
+> 记录状态核对命令、审查范围、关键结果和未覆盖部分。
 > 按 `.agents/rules/evidence-reporting.md` 同时记录审查范围、关键结果和未覆盖部分；正常成功不粘贴完整 stdout。
 
 ## 证据原文
 
-> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对 `$ ` 命令和相称结果摘要；Blocker、失败、阻塞或争议必须保留可复现证据、准确位置和决定性摘录；无法复现的判断须降级或移入「自我质疑」。
+> 遵循 `.agents/rules/evidence-reporting.md`：每条断言配对可执行命令和相称结果摘要；Blocker、失败、阻塞或争议必须保留可复现证据、准确位置和决定性摘录；无法复现的判断须降级或移入「自我质疑」。
 
 - 断言：{verified claim}
 ```text
-$ {command}
+{command}
 {result summary or decisive excerpt}
 ```