webnovel-review

Project Root Guard(必须先确认)

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "webnovel-review" with this command: npx skills add lingfengqaq/webnovel-writer/lingfengqaq-webnovel-writer-webnovel-review

Quality Review Skill

Project Root Guard(必须先确认)

  • Claude Code 的“工作区根目录”不一定等于“书项目根目录”。常见结构:工作区为 D:\wk\xiaoshuo ,书项目为 D:\wk\xiaoshuo\凡人资本论 。

  • 必须先解析真实书项目根(必须包含 .webnovel/state.json ),后续所有读写路径都以该目录为准。

环境设置(bash 命令执行前):

export WORKSPACE_ROOT="${CLAUDE_PROJECT_DIR:-$PWD}"

if [ -z "${CLAUDE_PLUGIN_ROOT}" ] || [ ! -d "${CLAUDE_PLUGIN_ROOT}/skills/webnovel-review" ]; then echo "ERROR: 未设置 CLAUDE_PLUGIN_ROOT 或缺少目录: ${CLAUDE_PLUGIN_ROOT}/skills/webnovel-review" >&2 exit 1 fi export SKILL_ROOT="${CLAUDE_PLUGIN_ROOT}/skills/webnovel-review"

if [ -z "${CLAUDE_PLUGIN_ROOT}" ] || [ ! -d "${CLAUDE_PLUGIN_ROOT}/scripts" ]; then echo "ERROR: 未设置 CLAUDE_PLUGIN_ROOT 或缺少目录: ${CLAUDE_PLUGIN_ROOT}/scripts" >&2 exit 1 fi export SCRIPTS_DIR="${CLAUDE_PLUGIN_ROOT}/scripts"

export PROJECT_ROOT="$(python "${SCRIPTS_DIR}/webnovel.py" --project-root "${WORKSPACE_ROOT}" where)"

0.5 工作流断点(best-effort,不得阻断主流程)

目标:让 /webnovel-resume 能基于真实断点恢复。即使 workflow_manager 出错,也只记录警告,审查继续。

推荐(bash):

python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow start-task --command webnovel-review --chapter {end} || true

Step 映射(必须与 workflow_manager.py get_pending_steps("webnovel-review") 对齐):

  • Step 1:加载参考

  • Step 2:加载项目状态

  • Step 3:并行调用检查员

  • Step 4:生成审查报告

  • Step 5:保存审查指标到 index.db

  • Step 6:写回审查记录到 state.json

  • Step 7:处理关键问题(AskUserQuestion)

  • Step 8:收尾(完成任务)

Step 记录模板(bash,失败不阻断):

python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow start-step --step-id "Step 1" --step-name "加载参考" || true python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow complete-step --step-id "Step 1" --artifacts '{"ok":true}' || true

Review depth

  • Core (default): consistency / continuity / ooc / reader-pull

  • Full (关键章/用户要求): core + high-point + pacing

Step 1: 加载参考(按需)

References(按步骤导航)

  • Step 1(必读,硬约束):core-constraints.md

  • Step 1(可选,Full 或节奏/爽点相关问题):cool-points-guide.md

  • Step 1(可选,Full 或节奏/爽点相关问题):strand-weave-pattern.md

  • Step 1(可选,仅在返工建议需要时):common-mistakes.md

  • Step 1(可选,仅在返工建议需要时):pacing-control.md

Reference Loading Levels (strict, lazy)

  • L0: 先确定审查深度(Core / Full),再加载参考。

  • L1: 只加载 References 区的“必读”条目。

  • L2: 仅在问题定位需要时加载 References 区的“可选”条目。

必读:

cat "${SKILL_ROOT}/../../references/shared/core-constraints.md"

建议(Full 或需要时):

cat "${SKILL_ROOT}/../../references/shared/cool-points-guide.md" cat "${SKILL_ROOT}/../../references/shared/strand-weave-pattern.md"

可选:

cat "${SKILL_ROOT}/references/common-mistakes.md" cat "${SKILL_ROOT}/references/pacing-control.md"

Step 2: 加载项目状态(若存在)

cat "$PROJECT_ROOT/.webnovel/state.json"

Step 3: 并行调用检查员(Task)

调用约束:

  • 必须通过 Task 工具调用审查 subagent,禁止主流程直接内联审查结论。

  • 各 subagent 结果全部返回后再生成总评与优先级。

Core:

  • consistency-checker

  • continuity-checker

  • ooc-checker

  • reader-pull-checker

Full 追加:

  • high-point-checker

  • pacing-checker

Step 4: 生成审查报告

保存到:审查报告/第{start}-{end}章审查报告.md

报告结构(精简版):

第 {start}-{end} 章质量审查报告

综合评分

  • 爽点密度 / 设定一致性 / 节奏控制 / 人物塑造 / 连贯性 / 追读力
  • 总评与等级

修改优先级

  • 🔴 高优先级(必须修改)
  • 🟠 中优先级(建议修改)
  • 🟡 低优先级(可选优化)

改进建议

  • 可执行的修复建议

审查指标 JSON(用于趋势统计):

{ "start_chapter": {start}, "end_chapter": {end}, "overall_score": 48, "dimension_scores": { "爽点密度": 8, "设定一致性": 7, "节奏控制": 7, "人物塑造": 8, "连贯性": 9, "追读力": 9 }, "severity_counts": {"critical": 1, "high": 2, "medium": 3, "low": 1}, "critical_issues": ["设定自相矛盾"], "report_file": "审查报告/第{start}-{end}章审查报告.md", "notes": "" }

注意:此处只生成审查指标 JSON;落库见 Step 5。

Step 5: 保存审查指标到 index.db(必做)

python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" index save-review-metrics --data '@review_metrics.json'

Step 6: 写回审查记录到 state.json(必做)

将审查报告记录写回 state.json.review_checkpoints ,用于后续追踪与回溯(依赖 update_state.py --add-review ):

python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" update-state -- --add-review "{start}-{end}" "审查报告/第{start}-{end}章审查报告.md"

Step 7: 处理关键问题

如发现 critical 问题(severity_counts.critical > 0 或 critical_issues 非空),必须使用 AskUserQuestion 询问用户:

  • A) 立即修复(推荐)

  • B) 仅保存报告,稍后处理

若用户选择 A:

  • 输出“返工清单”(逐条 critical 问题 → 定位 → 最小修复动作 → 注意事项)

  • 如用户明确授权可直接修改正文文件,则用 Edit 对对应章节文件做最小修复,并建议重新运行一次 /webnovel-review 验证

若用户选择 B:

  • 不做正文修改,仅保留审查报告与指标记录,结束本次审查

Step 8: 收尾(完成任务)

python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow start-step --step-id "Step 8" --step-name "收尾" || true python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow complete-step --step-id "Step 8" --artifacts '{"ok":true}' || true python "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" workflow complete-task --artifacts '{"ok":true}' || true

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

webnovel-write

No summary provided by upstream source.

Repository SourceNeeds Review
General

webnovel-plan

No summary provided by upstream source.

Repository SourceNeeds Review
General

webnovel-init

No summary provided by upstream source.

Repository SourceNeeds Review
General

webnovel-resume

No summary provided by upstream source.

Repository SourceNeeds Review