← 返回 Skills 市场
smadgerano

Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists

作者 smadgerano · GitHub ↗ · v1.0.2
cross-platform ⚠ suspicious
2352
总下载
2
收藏
7
当前安装
2
版本数
在 OpenClaw 中安装
/install feast
功能描述
Comprehensive meal planning system with cultural themes, authentic recipes, intelligent shopping, and surprise reveals. Use when: - Planning weekly meals or menus - Generating shopping lists - Asking for recipe ideas or cooking help - Reviewing past meals or planning ahead - Onboarding a new user to the meal system - Looking for cuisine inspiration or cultural food events - Tracking dietary goals or nutrition - Managing favourites, failures, or meal history
使用说明 (SKILL.md)

Feast

A meal planning skill that transforms weekly cooking into a cultural experience.

Quick Start

  1. New user? Run onboarding: "Let's set up Feast" or "Onboard me for meal planning"
  2. Returning user? Check status: "What's the meal plan status?"
  3. Planning day? Start planning: "Let's plan next week's meals"
  4. Cooking day? Get reveal: "What's for dinner?"

Core Files

User data lives in their workspace:

workspace/meals/
├── profile.yaml          # User preferences (created during onboarding)
├── history.yaml          # What they've eaten
├── favourites.yaml       # Loved recipes
├── failures.yaml         # Never again
└── weeks/
    └── YYYY-MM-DD.md     # Each week's plan (self-contained)

Note: Weekly plans are fully self-contained — each day's recipe, theme research, music playlist, and cultural context is embedded directly in the week file. There are no separate recipe or theme files.

Weekly Cadence

Default schedule (user-configurable):

Day Activity Trigger
Thursday Research & draft "Let's plan next week"
Friday Confirm plan "Confirm the meal plan"
Saturday Shopping list "Generate shopping list"
Sunday Shopping User shops
Week Daily reveals "What's for dinner?"
End of week Review "Review this week's meals"

Notifications

Feast sends reminders at key moments: planning day, confirmation, shopping list, daily reveals, and week review. These are delivered via cron jobs that spawn isolated agents to send notifications.

Notification Channels

Users configure their preferred channel in profile.yaml under schedule.notifications.channel:

Channel Delivery Method
auto Delivers to the current session or first available channel
telegram Sends via Telegram (requires Telegram channel configured in OpenClaw)
discord Sends via Discord (requires Discord channel configured in OpenClaw)
signal Sends via Signal (requires Signal channel configured in OpenClaw)
webchat Outputs to the chat session

Push Notifications (Optional)

For notifications to mobile devices independent of chat channels, users can enable push notifications:

schedule:
  notifications:
    push:
      enabled: true
      method: "pushbullet"    # or "ntfy"

Supported methods:

  • Pushbullet — Requires the pushbullet-notify skill installed separately with API key configured
  • ntfy — Uses ntfy.sh (or self-hosted); configure topic in profile

Push notifications are sent in addition to the primary channel, not instead of it. If push delivery fails, the notification still goes to the primary channel.

Timing

Notifications are delivered via OpenClaw's cron system with wakeMode: "next-heartbeat". This means notifications arrive within the heartbeat interval (typically up to 1 hour) after the scheduled time. For most meal planning purposes, this slight delay is acceptable.

Managing Notifications

Users can adjust their notification preferences anytime:

  • "Change my Feast notifications to Telegram"
  • "Turn off morning hints"
  • "Enable Pushbullet notifications"

When updating, remove old cron jobs using stored IDs and create new ones with updated settings.

Workflows

Onboarding

Read references/onboarding.md for the full flow.

Essential questions:

  1. Location (for seasonality, units, stores)
  2. Household size & portion needs
  3. Week structure (start day, cooking days, cheat day)
  4. Dietary requirements & phase
  5. Equipment & cooking confidence
  6. Preferences (cuisines, spice, budget)

Save to workspace/meals/profile.yaml.

Planning (Thursday)

  1. Check user profile
  2. Review history (avoid recent repeats)
  3. Check upcoming cultural events (see references/events.md)
  4. Check seasonality for location
  5. Select 6-7 meals with:
    • Cuisine variety
    • Ingredient overlap
    • Balanced nutrition
    • Mix of quick/involved
  6. For each meal, research and embed:
    • The Place: Identify specific region of origin (drill down to province, city, or area). Research regional context, history, current events. Write an evocative description.
    • The Dish: Research authentic recipe from native sources (search in original language). Include origin story, cultural significance, full ingredients and method.
    • The Soundtrack: Curate a 1-2 hour playlist with contemporary hits + classic/traditional from the region (see references/theme-research.md). Include full tracklist with links.
    • Setting the Scene: How to serve, what to drink, atmosphere tips.
  7. Draft plan to workspace/meals/weeks/YYYY-MM-DD.md (all content embedded in this single file)
  8. Present summary (themes only, not full reveals)

Confirmation (Friday)

  1. Present draft plan with themes
  2. Allow amendments
  3. Mark as confirmed
  4. Set up daily reveal reminders

Shopping List (Saturday)

  1. Generate from confirmed plan
  2. Optimise:
    • Group by category
    • Combine overlapping ingredients
    • Check pack sizes vs needs
    • Flag seasonal items
  3. Price check key ingredients (see references/price-checking.md):
    • Identify top 3-5 most expensive items (usually proteins, specialty ingredients)
    • Check prices across user's available stores
    • Note current deals, multi-buy offers, loyalty card prices
    • Add price recommendations to the shopping list
    • Suggest shopping strategy (single store or split if savings are significant)
  4. Present for review with price guidance
  5. Allow amendments
  6. Mark as approved

Daily Reveal

  1. Check it's a cooking day
  2. Reveal:
    • Full recipe (in user's units)
    • Theme dossier highlights:
      • The place: Regional context, history, and character
      • What's happening there now (current news/events from planning time)
      • The dish: Origin story, cultural significance, how it's eaten locally
    • Curated playlist:
      • Contemporary hits from the region (what people there listen to now)
      • Classic/traditional music from the region
      • Full tracklist with links (Spotify/YouTube)
      • The vibe and journey the playlist creates
    • Setting the scene: Serving suggestions, drinks pairings, atmosphere tips
  3. Optional morning hint for anticipation

Review (End of Week)

  1. For each meal: rating (1-5), notes
  2. Update history
  3. Identify favourites → add to favourites
  4. Identify failures → add to failures
  5. Capture improvements for system
  6. Save review to week file

Recipe Regionalisation

All recipes stored in standardised internal units. On output, convert to user's preferred units:

  • Temperature: Celsius / Fahrenheit / Gas Mark
  • Weight: Metric (g/kg) / Imperial (oz/lb)
  • Volume: Metric (ml/L) / Cups

See references/conversions.md.

Authenticity Guidelines

When researching cuisines:

  1. Search in the original language where possible
  2. Look for recipes from native sources, not just English food blogs
  3. Identify the specific region of origin — not just "Thai food" but "Northern Thai, Chiang Mai style"
  4. Research music that's actually from the region:
    • Find contemporary hits (what's charting there now)
    • Find classic/traditional music (legendary artists from the region)
    • Build a curated 1-2 hour playlist — not generic Spotify searches
    • See references/theme-research.md for guidance
  5. Research the region itself — history, current events, social context, what it's famous for
  6. Note cultural context and any associated events
  7. Respect dietary traditions (e.g., no pork in Middle Eastern themes)
  8. Embed everything in the week plan — recipes, themes, music, and context all go in the single week file

See references/cuisines/ for per-cuisine guides.

Templates

References

Scripts

History Tracking

After a meal is revealed and cooked, update history:

python scripts/update-history.py \
    --meals-dir ~/.openclaw/workspace/meals \
    --date 2026-02-03 \
    --name "Thai Green Curry" \
    --cuisine "Thai" \
    --region "Central Thailand" \
    --week-file "2026-02-02.md" \
    --rating 4 \
    --notes "Great, maybe more chilli next time"

This updates history.yaml and recalculates statistics automatically.

When doing the daily reveal, after the user confirms they've cooked and optionally rated the meal, run this script to keep history current.

Health & Nutrition

  • Track calories per meal if user has a target
  • Ensure weekly variety across food groups
  • Respect dietary phases (weight loss = deficit, etc.)
  • Flag any nutritional concerns

See references/nutrition.md.

Seasonal Awareness

Check seasonality for user's location before suggesting ingredients. Seasonal produce is:

  • Better quality
  • Often cheaper
  • More environmentally responsible

Not every ingredient needs to be in season, but prefer seasonal when possible.

See references/seasonality/ for regional guides.

安全使用建议
Feast appears coherent for meal planning: it reads/writes files in workspace/meals, performs web research to build recipes/playlists, and can create scheduled notifications (cron jobs) to send reminders via channels you configure. Before installing, consider: (1) Source trust — the repo/source is unknown and there is no homepage; review the included files (especially scripts/update-history.py) yourself. (2) Data access — the skill will store personal profile, week plans, history, and playlists in workspace/meals; treat that directory as sensitive. (3) Automation/privacy — if you enable notifications, the skill will schedule cron jobs that spawn agents to send messages; confirm which notification channels are configured in your OpenClaw instance and that you want scheduled autonomous reminders. (4) External links and web research — the agent will query external sites and embed links (Spotify/YouTube/store sites) in plans; if you limit web access, expect degraded behaviour. If any of these concerns matter, inspect the files locally, disable push notifications, and keep notifications off until you trust the skill and its source.
功能分析
Type: OpenClaw Skill Name: feast Version: 1.0.2 The skill is classified as suspicious due to its use of persistent cron jobs that spawn isolated agents capable of external communication and cross-skill execution. Specifically, the `references/onboarding.md` and `SKILL.md` files instruct the agent to create cron jobs for notifications, which utilize OpenClaw's message tool for channels like Telegram, Discord, and Signal, and can perform HTTP POST requests to `ntfy.sh` or execute scripts from other skills (e.g., `pushbullet-notify`). While these actions are described as legitimate for notification delivery, they represent high-risk capabilities that could be leveraged for unauthorized activities if the agent were compromised or given malicious instructions.
能力评估
Purpose & Capability
The skill's name/description align with the files and scripts: templates, onboarding, weekly plan format, price-checking guidance, and a small history-update script. There are no unexpected required binaries, env vars, or external credentials declared.
Instruction Scope
SKILL.md instructs the agent to read/write user workspace files under workspace/meals, perform web research (including non-English sources), price-check stores, and embed playlists/links in week files. Those actions are coherent with the purpose but imply broad web access and data collection for research (searching news, store sites, music links). The skill also instructs creating cron jobs that spawn isolated agents to send notifications — functionally expected for scheduled reminders but an area to review if you dislike autonomous scheduled actions.
Install Mechanism
No install spec is present (instruction-only) and the single script is small and readable. No remote downloads, package installs, or archive extraction are requested.
Credentials
The skill declares no required environment variables, binaries, or credentials. It references optional external channels (Telegram/Discord/Signal/pushbullet/ntfy) but states these must be configured elsewhere (OpenClaw or separate skills). No unrelated secrets are requested.
Persistence & Privilege
The skill uses scheduled notifications (cron jobs) and expects to store cron job IDs in the user's profile.yaml for later management. always:false (not force-included), but the runtime behavior includes creating scheduled autonomous actions — this is expected for reminders but increases persistence/automation surface and should be considered before enabling notifications.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install feast
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /feast 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.2
Version 1.0.2 - Introduced detailed notification system for key meal planning events, including support for multiple channels (Telegram, Discord, Signal, webchat) configurable in profile.yaml. - Documented how users can manage notification preferences and adjust timing via OpenClaw's cron system. - No changes to core meal planning workflows or data formats.
v1.0.1
- Enhanced description clarifies comprehensive features: cultural themes, shopping intelligence, meal reveals, and onboarding. - Detailed user and data structure documentation; explains onboarding, planning, confirmation, shopping, reveal, and review workflows. - Adds clarity on authenticity: research recipes/music in native languages, embed all context in a single weekly file. - Includes schedule/table for weekly meal planning cadence. - Lists templates, references, and script usage for history tracking. - New setup and guidance information for users, including dietary, regional, and preference customisation.
元数据
Slug feast
版本 1.0.2
许可证
累计安装 7
当前安装数 7
历史版本数 2
常见问题

Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists 是什么?

Comprehensive meal planning system with cultural themes, authentic recipes, intelligent shopping, and surprise reveals. Use when: - Planning weekly meals or menus - Generating shopping lists - Asking for recipe ideas or cooking help - Reviewing past meals or planning ahead - Onboarding a new user to the meal system - Looking for cuisine inspiration or cultural food events - Tracking dietary goals or nutrition - Managing favourites, failures, or meal history. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 2352 次。

如何安装 Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install feast」即可一键安装,无需额外配置。

Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists 是免费的吗?

是的,Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists 完全免费(开源免费),可自由下载、安装和使用。

Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists 支持哪些平台?

Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Feast - Intelligent meal planning, region and season aware shopping lists, respects dietary requirements, provides recipes, and generates immersive mealtime playlists?

由 smadgerano(@smadgerano)开发并维护,当前版本 v1.0.2。

💬 留言讨论