SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillswebnovel-query

webnovel-query

Writing & research

View repository

Updated 10-15 09:39

About this skill

Information Query Skill

Use when

Triggered when the user asks about in-project information such as story settings, characters, power systems, factions, foreshadowing, golden finger, pacing, etc.

Project root protection

export WORKSPACE_ROOT="${CLAUDE_PROJECT_DIR:-$PWD}"
export SCRIPTS_DIR="${CLAUDE_PLUGIN_ROOT}/scripts"
export SKILL_ROOT="${CLAUDE_PLUGIN_ROOT}/skills/webnovel-query"
export PROJECT_ROOT="$(python "${SCRIPTS_DIR}/webnovel.py" --project-root "${WORKSPACE_ROOT}" where)"
  • PROJECT_ROOT Must include .webnovel/state.json
  • ForbiddenUnder ${CLAUDE_PLUGIN_ROOT}/ read or write project files

Query category → Narrowest tool

First identify the query type, then use the narrowest tool in the table below. Do not load everything by default; only use it for comprehensive / cross-multiple-type queries memory-contract load-context。

Query type Keywords Narrowest tool
Character historical state A character at chapter N / state at a point in time / realm changes knowledge query-entity-state
Entity relationships Relationship / friend or foe / master-disciple / faction affiliation knowledge query-relationships
World rules Power rules / setting iron laws / realm system constraints memory-contract query-rules
Foreshadowing / open loop Foreshadowing / urgent foreshadowing / unresolved suspense memory-contract get-open-loops
Comprehensive / complex Cross-multiple-type, requiring timeline + long-term memory combined memory-contract load-context
Static settings Character cards / power systems / worldbuilding / factions / tag format Grep + Read Settings collection

Reference loading strategy

Load on demand by query type; identify first, then load. Path description:references/ refers to skill-private skills/webnovel-query/references/；../../references/ refers to shared references.

Query type Reference Actual path
Data flow / priority Data flow specification ${SKILL_ROOT}/references/system-data-flow.md
Foreshadowing analysis Foreshadowing analysis ${SKILL_ROOT}/references/advanced/foreshadowing.md
Pacing analysis Strand mode ${SKILL_ROOT}/../../references/shared/strand-weave-pattern.md
Format query Tag specification ${SKILL_ROOT}/references/tag-specification.md

Do not load more than two references at the same time unless the user's request explicitly spans multiple types.

Query process

  1. Identify query type: match keywords against the "Query category → Narrowest tool" table.

  2. Locate the pre-writing source of truth by priority(pre-writing source of truth → post-writing source of truth → projection layer):

    1. .story-system/MASTER_SETTING.json - Book-wide main settings (genre, tone, core taboos)
    2. .story-system/volumes/*.json - Volume-level contract (this volume's goals, pacing strategy)
    3. .story-system/chapters/*.json - Chapter-level contract (this chapter's focus, dynamic context)
    4. latest accepted .story-system/commits/chapter_XXX.commit.json - Post-writing facts (finalized state of published chapters)
    5. memory-contract series queries - memory orchestration results (long-term memory, foreshadowing, timeline)
    6. .webnovel/state.json / index.db - Projection layer (fallback / read-model only, analogous to the "character cards" and "chapter lists" in a web novel backend)

    Priority description：

    • Pre-writing source of truth (1-3): the "outline, settings, restricted zones" the author must follow before starting to write
    • Post-writing source of truth (4): the "finalized state" of published chapters, which cannot be tampered with
    • Projection layer (5-6): a "query view" automatically generated from the post-writing source of truth, for quick retrieval

  3. Call the narrowest tool for retrieval: call only the required command by type, do not default to loading everything load-context。

# 角色历史状态：某实体在指定章节时的状态
python -X utf8 "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" knowledge query-entity-state --entity "{entity_id}" --at-chapter {N}

# 实体关系：某实体在指定章节时的所有关系
python -X utf8 "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" knowledge query-relationships --entity "{entity_id}" --at-chapter {N}

# 世界规则
python -X utf8 "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" memory-contract query-rules --chapter {chapter_num}

# 伏笔 / open loop
python -X utf8 "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" memory-contract get-open-loops

# 仅综合 / 复杂查询：需要时间线 + 长期记忆联合时才用
python -X utf8 "${SCRIPTS_DIR}/webnovel.py" --project-root "${PROJECT_ROOT}" memory-contract load-context --chapter {chapter_num}

For static settings (character cards / power systems / worldbuilding / tag format), directly use Grep to locate the line number, then Read Take the fragment, without going through memory-contract.

  1. Formatted output: output according to the template below.

Output format

# 查询结果：{关键词}

## 概要
- **匹配类型**: {type}
- **数据源**: {实际命中的真源 / 投影层}
- **匹配数量**: X 条

## 详细信息
{结构化数据，含文件路径和行号}

## 数据一致性检查
{state.json 与静态文件的差异，若无差异则省略}

Boundaries and failure recovery

  • Read-only operation, does not modify any project files
  • If the data source is missing, clearly inform the user which file is missing
  • If the query has no match, return an empty result and suggest checking the scope
  • If .story-system/ contract and accepted commit are missing, you must explicitly state that the current query has been downgraded to legacy fallback