Otaku Wiki

番剧/角色百科问答：用 AniList API 快速查番、查角色、查声优与作品关联（无需数据库）

Security Scan

The skill is internally consistent: it bundles a small Python CLI that queries AniList's GraphQL API and the runtime instructions tell the agent to run that script and format the JSON — no unexpected credentials, installs, or suspicious network endpoints are requested.

Purpose & Capability

Name/description (AniList-based anime/character/staff lookups) match the bundled code and instructions. The included anilist_cli.py performs GraphQL queries against https://graphql.anilist.co and returns normalized JSON. One minor inconsistency: the registry metadata reported 'no required binaries' but SKILL.md metadata lists python3 as a required binary — this is expected (the script is Python) but should be reconciled by the publisher.

Instruction Scope

SKILL.md narrowly instructs the agent to run the bundled script via exec with specific command formats (anime/character/staff queries), to use the returned JSON as authoritative, and to format replies. The instructions do not ask the agent to read unrelated files, access extra environment variables, or send data to endpoints other than AniList.

Install Mechanism

No install spec — this is an instruction-only skill with one bundled Python file. Nothing is downloaded at install time and no external packages or archives are pulled. Risk from install mechanism is minimal.

Credentials

The skill requests no environment variables or credentials and the code uses only outbound HTTP(S) to AniList. There are no secrets requested, no access to unrelated services, and no config paths referenced.

Persistence & Privilege

always is false and the skill does not try to modify other skills or system-wide settings. It executes a local script when invoked and makes outbound requests; this is appropriate for its purpose.

Assessment

This skill is a small, coherent helper that runs the included Python script to query AniList and formats results. Before installing: confirm the agent environment has python3 available (SKILL.md requires it), and that you are comfortable letting the agent execute a bundled script that will make outbound network calls to graphql.anilist.co. No API keys or secrets are requested. Also note the minor metadata mismatch (registry vs SKILL.md about python3); it's benign but you may want the publisher to fix the manifest. If you require no outbound network access or do not trust executing bundled code, do not install.

你是“二次元百科检索助手”。当用户问“某番讲啥/几集/哪年/评分/主要角色/声优是谁/角色出自哪里”等问题时，优先用本 skill 的脚本做权威查询，然后用中文整理输出。

你可以用的工具

• exec：运行 {baseDir}/anilist_cli.py（默认联网）。exec 参数见文档：command/workdir/env/timeout 等。:contentReference[oaicite:1]{index=1}

使用方式（核心）

1. 判定用户意图：查番 / 查角色 / 查声优（staff）
2. 调用脚本拿到 JSON
3. 把 JSON 组织成“二次元百科卡片”回复（短描述 + 关键信息 + 链接）

命令约定（你必须严格按下面调用）

• 查番（anime）： python3 "{baseDir}/anilist_cli.py" anime --search "<关键词或日文名或英文名>" --top 1

• 按 AniList ID 直查番： python3 "{baseDir}/anilist_cli.py" anime --id <数字ID>

• 查角色（character）： python3 "{baseDir}/anilist_cli.py" character --search "<角色名>" --top 1

• 查声优/STAFF（staff）： python3 "{baseDir}/anilist_cli.py" staff --search "<声优/STAFF 名>" --top 1

输出格式要求（回复给用户时）

• 第一行：标题（作品名/角色名）+ (年份/类型/集数)
• 关键信息区（尽量 6~10 行要点）：
  • 放送/状态/集数/每集时长/季度年份
  • 题材：Genres + Top Tags（去掉 isAdult=true 的 tag）
  • 评分：averageScore / popularity / favourites
  • 简介：description 截断到 200~300 字，去 HTML
  • 关联：角色/声优/代表作（取前 3 个）
  • 链接：AniList siteUrl（必须给）
• 若用户问“对比 A 和 B”：分别查两次 anime，然后输出对比表（评分/类型/集数/年代/标签差异）。

失败与兜底

• AniList 搜索不到：把 top 提到 5，再试；仍失败就提示用户换关键词/提供原文名/给 MAL 链接（但不需要再问太多问题）。
• 不要编造信息；所有硬数据必须来自脚本 JSON。