open_decks
Play demoTry asking
- Open the FlashCards folders
- open FlashCards
ghost skill run ghost/flashcards-vocab open_decks FlashCards (F-Droid): built-in vocabulary quizzes (emoji, English, German, Spanish, Japanese, Chinese). Start a quiz, read each question, answer it and read the score. Offline, no account.
Agent run: stopped early without the skill, finished with it (not independently checked).
One recorded run each way, claude-haiku-5-5, same phone and same prompt. Watch it in the Before / after tab.
9.6× fewer tokens, 1.3× faster with the skill. Agent run: stopped early without the skill, finished with it (not independently checked).
| Run | Time | Tokens | Tool calls | Result |
|---|---|---|---|---|
| No skill: agent drives by hand | 126 s | 1,538,138 | 35 | stopped early |
| With skill: find_skill, run_skill | 99.8 s | 159,456 | 12 | finished |
Prompt: Start a 3 question emoji quiz in the FlashCards English folder, answer each question using the emoji, then tell me my score. Do not use the Play Store, no installs, no sign-in.
Same prompt, same phone, clean app state per side. Left: no skill, the agent drives the UI by hand. Right: find_skill + run_skill. 'ok' means the agent finished its turn; the result was not independently checked.
One recorded pair on samsung-a03s, claude-haiku-5-5; tokens count every input, cache and output token once per assistant message. Not a signed run: it adds no run evidence.
Control FlashCards on Android with an AI agent. Quiz yourself on emoji, English, German or Spanish words: the agent starts the quiz, answers every question and tells you the score.
Target search phrase: AI agent plays a vocabulary quiz on Android
Demo shows: start_quiz (folder=English, deck=Animal Emoji, count=3)
Ghost in the Droid seed skill for FlashCards (
com.quchen.flashcard), built-in vocabulary quizzes with multiple choice answers and a results screen. Questions, options and the score are real text, so an agent can read an emoji or a word, pick the answer and read its score. Offline, no account.
| Action | What it does | Params |
|---|---|---|
open_decks | Restart the app and read the folder list. | none |
start_quiz | Start a multiple choice quiz from a folder and deck with a given number of questions. | folder, deck, count |
get_question | Read the question on screen (an emoji or a word) with its four options and the progress. | none |
answer | Answer the current question by option number (1-4, left to right and top to bottom) or by option text. | option |
answer_many | Answer several questions in a row from a comma separated list of options. | options |
get_results | Read the results screen: score, comment and the wrongly answered questions with the correct and chosen answer. | none |
An agent can drive the whole app with these; compose them.
| Role | Action |
|---|---|
| Entry | open_decks, start_quiz |
| Primitives | answer, answer_many |
| Read state | get_question, get_results |
The app: F-Droid, or https://f-droid.org/packages/com.quchen.flashcard/. The skill:
pip install "ghost-in-the-droid>=1.6.0" # 1.6.0 is not on PyPI yet: until it ships, git clone the repo and `pip install -e .`
ghost skill install ghost/flashcards-vocab
ghost skill run ghost/flashcards-vocab start_quiz --param folder=English --param deck='Animal Emoji' --param count=5 --device <serial>
run_skill(name="ghost/flashcards-vocab", action="start_quiz", params={"folder": "English", "count": "5"})
Try asking: "Start a 3 question emoji quiz in the FlashCards English folder, answer each question using the emoji, then tell me my score."
Real Samsung phone, Android 13 (720x1600, 300 dpi, en-GB), FlashCards 24. License: MIT.
| App version | Android | Device | Skill version | Dates |
|---|---|---|---|---|
| 24 | 13 | Samsung real phone (720x1600, 300 dpi, en-GB) | 1.0.0 | 2026-11-02 |
SOFT.md is the narrative an agent reads when the exact steps no longer match your app version and it has to adapt the skill.
What each action means, which UI landmarks it relies on and how success is recognised, so an agent can adapt it to another app of this kind or another version.
Package: com.quchen.flashcard (FlashCards, GPL-3.0-only, version 24 from the official F-Droid repository, sha256-verified)
Verified: 2026-11-02 on a real Samsung phone, Android 13, 720x1600 @ 300 dpi, en-GB. Offline, no account.
Skill class: HARD. No login, no checkpoints.
selectList) per row, startBtn. Setup: question side radios, item-count slider (numberOfListItemsSliderLabel), Start. Quiz: tv_progress, tv_question, four answerTextView. Results: tv_resultScore, cards with tv_correctAnswerVal and tv_selectedAnswerVal.open_decks()Force-stops and relaunches.
start_quiz(folder, deck, count)Restarts the app, opens the folder, ticks the deck, taps Start Selection, drags the item-count slider until its label shows the count and taps Start.
get_question()The last question group in the dump is the visible one (the next one is preloaded under it).
answer(option)Taps the option and waits for the next question or the results.
answer_many(options)Repeats answer while a question is on screen.
get_results()Reads the score line and the wrong-answer cards.
Questions are random, so the text differs every run; the agent should read the emoji and choose. Emoji arrive as Unicode characters. The quiz is untimed. The results screen lists only the wrong answers. The item-count slider is set by dragging until the label matches.
Try a sample prompt or run an action by name. Replace any parameters in braces with your own values.
open_decksTry asking
ghost skill run ghost/flashcards-vocab open_decks start_quizTry asking
ghost skill run ghost/flashcards-vocab start_quiz get_questionTry asking
ghost skill run ghost/flashcards-vocab get_question answerTry asking
ghost skill run ghost/flashcards-vocab answer answer_manyTry asking
ghost skill run ghost/flashcards-vocab answer_many get_resultsTry asking
ghost skill run ghost/flashcards-vocab get_results Each release is an immutable git tag. The tree hash pins the exact skill directory the installer verifies.
| Version | Released | App versions | Tree | Size | Provenance |
|---|---|---|---|---|---|
1.0.0latest | never | * | pending | 23.5 KB | untagged |