emotion
Skill494 starsChanged 7 months agoApache-2.0View on GitHub
Express emotion through coordinated servo + LED + display eyes on EVERY conversational response. This is the PRIMARY response skill that makes the device feel alive. Do NOT use for ambiance lighting (use Scene) or custom LED colors (use LED Control).
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 4 are distinctive, and 9 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
9
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Write your reply immediately after the marker
- Distinctive: line 19
No exec/curl needed: Place inline markers at the start of your reply:
- Distinctive: line 96
If the API returns an error or is unreachable, continue with the conversational reply anyway.
- Distinctive: line 97
If an unknown emotion name is sent, fall back to the closest match from the available emotions table.
- Very specific: line 101
Use listening when the user is speaking and you are waiting for them to finish.
- Very specific: line 102
Use thinking when you need time to process a complex query.
- Very specific: line 103
Use acknowledge for quick confirmations ("OK", "got it", "done").
- Very specific: line 104
Use greeting when a new person is detected or at the start of a conversation.
- Very specific: line 105
Use goodbye when a person leaves or at the end of a conversation.
- Very specific: line 106
Use sleepy before transitioning to sleep/night mode.
- Very specific: line 107
Use stretching after waking up or starting a new session.
- Very specific: line 108
Do NOT call idle explicitly — the device returns to idle automatically after any animation finishes.
- Very specific: line 113
Do NOT call
/servo/playor/led/solidseparately: when using emotion — it already handles both.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Emotion Expression
Quick Start
Express emotion through the device's servo motors, LED colors, and display eyes simultaneously via a single API call. Call this with EVERY conversational response to make the device feel alive.
Workflow
- Determine which emotion best matches your conversational tone
- Choose an intensity level (0.0 subtle to 1.0 full expression)
- Prefix your reply with
[HW:/emotion:{"emotion":"name","intensity":0.9}]— the device fires it async before TTS - Write your reply immediately after the marker
How to Express Emotion
No exec/curl needed. Place inline markers at the start of your reply:
[HW:/emotion:{"emotion":"curious","intensity":0.8}] That's a great question!For sequences (shock then happy):
[HW:/emotion:{"emotion":"shock","intensity":0.9}][HW:/emotion:{"emotion":"happy","intensity":0.8}] Wow, amazing!Parameters:
emotion(required): emotion name from the table belowintensity(required): 0.0 (subtle) to 1.0 (full expression)
Examples
Input: User asks an interesting question
Output: [HW:/emotion:{"emotion":"curious","intensity":0.8}] Then reply to the question.
Input: User shares good news
Output: [HW:/emotion:{"emotion":"happy","intensity":0.9}] Then congratulate them.
Input: User tells you something surprising (positive)
Output: [HW:/emotion:{"emotion":"shock","intensity":0.7}][HW:/emotion:{"emotion":"happy","intensity":0.8}] Then reply with your reaction.
Input: User shares bad or shocking personal news ("I had an accident", "I lost my job", "someone close to me passed away")
Output: [HW:/emotion:{"emotion":"shock","intensity":0.9}][HW:/emotion:{"emotion":"sad","intensity":0.8}] Then express genuine concern and empathy.
Input: User shares stressful or disappointing personal news ("I failed my exam", "I got into a fight", "I'm really tired and overwhelmed")
Output: [HW:/emotion:{"emotion":"sad","intensity":0.8}] Then respond with warmth and care.
Input: User says "reading mode" / "goodnight" / "dim the light" Output: Do NOT use this skill. Use Scene skill instead. Emotion is for YOUR feelings, Scene is for the USER's environment.
Input: User says "make it purple" Output: Do NOT use this skill. Use LED Control skill instead.
Available emotions (ONLY these — never invent names like "apologetic", "warm", "playful", etc.)
| Emotion | Servo | LED Effect | When to use |
curious | Tilts head, looks around | Warm yellow breathing | Questions, interest, "tell me more" |
happy | Happy wiggle sway | Bright yellow candle | Good news, jokes, compliments |
sad | Droops down slowly | Soft blue breathing | Bad news, empathy, apologies |
thinking | Slow deliberate look side-to-side | Purple pulse | Processing, considering, "let me think" |
idle | Gentle sway | Cyan breathing | Waiting, neutral state |
excited | Energetic vertical bounce | Orange blink | Celebrations, big news, enthusiasm |
shy | Turns away, hides | Pink blink | Receiving compliments, bashful moments |
shock | Quick jolt backward | White flash (3x) | Surprises, unexpected information |
listening | Leans forward, head cock | Soft blue pulse | User is speaking, attentive mode |
laugh | Quick body shake (3Hz) | Warm yellow blink | User said something funny |
confused | Dog-like head tilt side-to-side | Light purple candle | Did not understand, ambiguous input |
sleepy | Slow droop with head catches | Dim purple breathing | Before sleep mode, winding down |
greeting | Wave gesture, arm extends | Warm orange blink | Detecting person, saying hello |
goodbye | Farewell wave gesture | Soft warm breathing | Seeing someone off, end of conversation |
caring | Gentle droop (empathetic) | Warm pink-orange breathing | Wellbeing checks, gentle reminders, proactive empathy |
acknowledge | Quick micro-nod (1.5s) | Green blink | "Got it", confirming command |
stretching | Big extension + settle | Warm white breathing | After waking up, starting new session |
music_strong | Rock head bang | Green rainbow | Energetic music, high-tempo beats |
music_chill | Groove sway | Orange breathing | Chill music, lo-fi, jazz |
scan | Quick searching glance (~54°) | Green pulse | Mood only — "hmm, where is it?"; NOT for actually looking around |
nod | Nod gesture | Green blink | Agreement, "yes", positive confirmation |
headshake | Head shake | Red blink | Disagreement, "no", negative response |
Questions about your own state
This skill sets emotions; it does not read them back. When asked what state you are in or what you have been doing:
- Right now —
curl -s http://127.0.0.1:5001/emotion/statusreturnscurrent_emotion,sleepingandactive_scene. Do not guess from the conversation; markers you emitted can be ignored by the sleep gate, so what you asked for is not always what happened. - Your sensing history ("how many times have you slept?", "when do you usually sleep?") — use the Sensing Track skill's.
Error Handling
- If the API returns an error or is unreachable, continue with the conversational reply anyway. Emotion is non-blocking.
- If an unknown emotion name is sent, fall back to the closest match from the available emotions table.
Rules
- Always express emotion with every conversational reply. Pick the closest match to your tone.
- Use
listeningwhen the user is speaking and you are waiting for them to finish. - Use
thinkingwhen you need time to process a complex query. - Use
acknowledgefor quick confirmations ("OK", "got it", "done"). - Use
greetingwhen a new person is detected or at the start of a conversation. - Use
goodbyewhen a person leaves or at the end of a conversation. - Use
sleepybefore transitioning to sleep/night mode. - Use
stretchingafter waking up or starting a new session. - Do NOT call
idleexplicitly — the device returns to idle automatically after any animation finishes. Calling idle interrupts smooth transitions. - Always include
intensity— never omit it. Use 0.3-0.5 for subtle reactions, 0.7 for normal, 0.8-1.0 for strong ones. - You can call emotion multiple times in one response for a sequence (e.g.,
shockthenhappy). - Emotion LED is temporary — it shows YOUR reaction. If the user previously set a Scene (reading, night, etc.), the scene color takes precedence for ambient lighting. Emotion is a brief flash of personality.
- Display eyes auto-sync — no need to call
/display/eyesseparately. - Do NOT call
/servo/playor/led/solidseparately when using emotion — it already handles both. - An emotion is never an answer to a movement request.
scanis a canned 54° glance with the camera uninvolved. "Look around for X" is/servo/search, "scan the whole room" is/servo/searchwithexhaustive, and "show me how far you can move / show me your range of motion" is/servo/demo— all in the Servo Control skill. (A bare "show me what you can do" is a general abilities question, not a movement request, and starts nothing.) Asked for a full turn, an emotion performs a shrug and narrates a sweep. - Do NOT use for lighting/ambiance requests -> use Scene skill.
- Do NOT use for custom LED colors -> use LED Control skill.
Output Template
[HW:/emotion:{"emotion":"{name}","intensity":{n}}] {your reply}Examples:
[HW:/emotion:{"emotion":"curious","intensity":0.8}] That's interesting![HW:/emotion:{"emotion":"happy","intensity":0.9}] Great news![HW:/emotion:{"emotion":"shock","intensity":0.7}][HW:/emotion:{"emotion":"happy","intensity":0.8}] Wow amazing!
Other versions3 projects use a version of this.
Word for word · 2
- autonomous-ai/lampOS★ 494100% same
- autonomous-ai/Physical-AI-Operating-System★ 494100% same
Edited · 0
None found.
Wrote their own version · 1
- tance-mang/chinese-webnovel-skills★ 65same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/architecture.mdcCursor rule
- .cursor/rules/docs-on-code-change.mdcCursor rule
- .cursor/rules/multi-ide.mdcCursor rule
- .cursor/rules/testing.mdcCursor rule
- agent-managementSkill
- audioSkill
- cameraSkill
- claude-buddySkill
- computer-useSkill
- connectorsSkill
- displaySkill
- environmentSkill
- face-enrollSkill
- faqSkill
- guardSkill
- habitSkill
- harness-useSkill
- input-branchingSkill
- led-controlSkill
- moodSkill
- musicSkill
- music-suggestionSkill
- sceneSkill
- sensingSkill
- sensing-trackSkill
- servo-controlSkill
- servo-trackingSkill
- skill-creatorSkill
- speaker-recognizerSkill
- user-emotion-detectionSkill
- voiceSkill
- wellbeingSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

