跳到正文
cccskills
免费在 GitHub 公开

gourmet-organizer

Turn menus, menu photos, and tasting observations into organized restaurant notes. Preserve sensory details, add grounded culinary context, compare relevant earlier experiences, and keep separate visit records when updating an existing note.

中文简介正在准备中。当前显示英语简介。

原文查看原文说明

Organize menus and visit notes with grounded tasting analysis and comparisons to earlier experiences. Use when the user shares a restaurant menu, wants a tasting write-up, or asks to merge a visit into Notes/Gourmet.md without overwriting earlier visits.

查看安装方法

适用场景

  • Turn menu photos and impressions into tasting notes
  • Add a restaurant visit to Notes/Gourmet.md
  • Compare multiple visits to the same restaurant
  • Record differences in aroma, texture, and finish

包含的文件(2)

  • SKILL.md5.3 KB
  • agents/openai.yaml320 B

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Gourmet Organizer

Use this skill when

Use this skill when the user:

  • shares a tasting menu, menu photo, or dish list;
  • asks for a restaurant tasting note or summary;
  • wants a new visit merged into Notes/Gourmet.md; or
  • wants multiple visits to the same restaurant kept side by side.

Workflow

1. Find the right landing spot

  1. Use the user-selected note, or Notes/Gourmet.md when it exists. Inspect headings with the available Markdown outline tool; missing prior notes do not block a new visit record.
  2. Search the chosen note for the restaurant and related cuisine sections. Read only relevant earlier visits for comparison.
  3. Prefer updating an existing restaurant section.
  4. If the restaurant already exists, preserve prior visits and split the notes by visit, date, or menu type rather than blending everything together.

1.5 Lightweight public research

Before writing the final note, do a small public lookup unless the user explicitly says not to:

  1. Search the public web for the restaurant name plus city/cuisine/dish keywords.
  2. Prefer official restaurant pages, Michelin/Black Pearl/Dianping/Google Maps/Apple Maps/Trip.com listings, local food media, and reliable cuisine references.
  3. Use public research only to confirm restaurant positioning, cuisine style, signature dishes, branch identity, and dish/cuisine background.
  4. Keep user-provided tasting notes as the source of truth for what was actually eaten and how it tasted.
  5. If sources are gated, noisy, or only partially readable, say so and avoid overclaiming.
  6. Add links only for sources actually used in the note; do not cite search snippets that were not opened or cannot be reasonably verified.

2. Normalize the source material

  1. Separate menu-provided information from user-added observations.
  2. If a menu photo is hard to read, use the photo only as a draft parse and treat explicit user corrections as the source of truth.
  3. Keep track of selected dishes versus menu options that were not chosen.
  4. Omit receipt-like details by default: price, quantity, serving count, and portion count. Keep them only when they materially affect the tasting note, such as an unusually large whole fish, a per-person tasting course, a paired comparison, or a value/portion judgment the user explicitly cares about.
  5. Preserve concrete sensory details: texture, aroma, temperature, aftertaste, and what made the dish memorable.

2.5 Develop the tasting analysis

  • By default, add useful tasting interpretation alongside the record. A complete dish list alone does not fulfill a tasting-note request. If the user explicitly wants only a list, respect that scope.
  • Start from distinctive dishes and the user's sensory words. Explain relevant ingredients, technique, temperature, aroma, texture, seasoning, or finish; choose the dimensions that teach something rather than applying a fixed template to every dish.
  • Read relevant earlier restaurant or ingredient notes and connect genuinely comparable experiences. Explain the basis of comparison; do not force a restaurant ranking or invent a connection when none is useful.
  • Distinguish actual tasting observations, sourced culinary background, and proposed things to notice next time. Use language such as “品鉴可看” or “可与……对照” for analysis that the user has not confirmed. Traditional recipes and menu names do not establish the restaurant's actual ingredients or process.
  • If public research is blocked, retain clearly qualified general tasting guidance where useful. Report the lookup limit; do not silently reduce the result to transcription or invent restaurant facts.

3. Write the note in the house style

  • Keep the existing restaurant heading if one already exists.
  • Use compact bullets and short tasting analysis instead of long generic prose.
  • Preserve the user's viewpoint implicitly through the tasting language; do not write meta phrases like "the user's note says".
  • Keep existing content intact. Add, refine, and structure, but do not delete valuable prior notes.
  • If the meal has a clear progression, keep the course order.
  • End with a short summary only when it helps compare visits or explain the restaurant's style.

4. Sources and safety

  • Add links only for external sources you actually used.
  • For user-provided menu photos or text, no external citation is needed.
  • Never overwrite a previous visit just because a newer menu is more detailed.

5. Check the actual result

Read the finished note: besides preserving dishes and sensory details, does it add a useful understanding of flavor or technique? Does any comparison to earlier notes explain a meaningful similarity or difference? Are unconfirmed ideas clearly phrased as tasting guidance? Fix these omissions before reporting completion; counts and formatting checks alone are insufficient.

Recommended output pattern

### [rating] Restaurant Name

#### First visit: menu name
- dish
  - ingredients
  - tasting note

#### Second visit: menu name
- dish
  - ingredients
  - tasting note

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Gather recent AI products, papers, tools, and viewpoints into a Chinese and English HTML daily report. Select up to ten supported items, check original sources where possible, and summarize shared trends with practical next actions.

中文简介正在准备中。当前显示英语简介。

  • Reviewing recent AI developments each morning
  • Tracking research and tools for AI agents
  • Choosing what to read or try next
huangruiteng/CS-Notes40112026年10月23日 更新

Retrieve an arxiv paper's full text from its HTML version using an ID or URL. Extract readable text with section structure, optionally save it to a file, and prepare it for reading, notes, or technical analysis.

中文简介正在准备中。当前显示英语简介。

  • Reading a paper from an arxiv ID or URL
  • Extracting full text for research notes
  • Saving long papers for section-by-section reading
huangruiteng/CS-Notes40112026年10月23日 更新

ask-echo

免费

Search the web or find images through the Ask-Echo API from 火山引擎. Retrieve result details, request an AI-generated summary, and narrow searches by time range or selected websites.

中文简介正在准备中。当前显示英语简介。

  • Finding web pages about a specific topic
  • Getting an AI-generated summary of search results
  • Searching for images
huangruiteng/CS-Notes40112026年10月23日 更新

Build or update the BlueBubbles channel plugin for Clawdbot. The skill explains code locations, existing helper functions, message delivery, connection checks, and webhook handling, which receives incoming messages and attachments through HTTP notifications.

中文简介正在准备中。当前显示英语简介。

  • Building or updating the BlueBubbles channel plugin
  • Implementing message delivery and connection checks
  • Handling incoming messages and attachments through webhooks
huangruiteng/CS-Notes40112026年10月23日 更新

Turn API documentation, SDKs, or existing scripts into reusable command-line tools for Codex. Build focused read and write commands, stable JSON output, authentication checks, and a companion skill explaining how future sessions should use the tool.

中文简介正在准备中。当前显示英语简介。

  • Turning an API into reusable commands
  • Making an internal script usable across projects
  • Adding JSON output and authentication checks
huangruiteng/CS-Notes40112026年10月23日 更新

Improve existing code through up to two optimization rounds, measuring execution time and memory use against a baseline. Verify that results remain correct, save the optimized code, and document changes, measured gains, and trade-offs in a report.

中文简介正在准备中。当前显示英语简介。

  • Improving the performance of existing code
  • Comparing execution time and memory use with a baseline
  • Checking that optimized code produces correct results
huangruiteng/CS-Notes40112026年10月23日 更新

查看 huangruiteng 的全部技能

报告此技能的问题