get_question
Play demoTry asking
- Show me the current kana quiz question
- what kana is the quiz asking
- start a Japanese kana quiz
ghost skill run ghost/kana-quiz get_question The Kana Quiz (F-Droid): read the kana question and its answer choices, answer, read the score. Offline, no account.
Agent run: finished 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.
3× fewer tokens, 1.3× slower with the skill. Agent run: finished without the skill, finished with it (not independently checked).
| Run | Time | Tokens | Tool calls | Result |
|---|---|---|---|---|
| No skill: agent drives by hand | 37.2 s | 161,617 | 12 | finished |
| With skill: find_skill, run_skill | 47.5 s | 53,615 | 7 | finished |
Prompt: Open the Kana Quiz on my phone and answer the first three questions correctly. 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 The Kana Quiz on Android with an AI agent. Practice Japanese on your phone: an agent reads the kana question, picks the romaji answer and tracks the score.
Target search phrase: AI agent Japanese kana quiz on Android
Demo shows: answer (choice=a)
Ghost in the Droid seed skill for The Kana Quiz (
com.noprestige.kanaquiz), an offline multiple-choice quiz that asks for the romaji of a Japanese kana character. Question, choices and score are all readable. Offline, no account.
| Action | What it does | Params |
|---|---|---|
get_question | Open the quiz; read the kana, the choices and the score | none |
answer | Tap the answer with this text; returns was_correct, score and the next question | choice |
get_score | Read correct, total and percent | none |
An agent can drive the whole app with these; compose them.
| Role | Action |
|---|---|
| Entry | get_question |
| Primitives | answer |
| Read state | get_score |
The app: F-Droid, or https://f-droid.org/packages/com.noprestige.kanaquiz/. 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/kana-quiz
ghost skill run ghost/kana-quiz get_question --device <serial>
ghost skill run ghost/kana-quiz answer --param choice=a --device <serial>
run_skill(name="ghost/kana-quiz", action="answer", params={"choice": "a"})
# {"result": {"picked": "a", "was_correct": true, "state": {"correct": 1, "total": 1, ...}}}
Try asking: "Open the Kana Quiz and answer the first three questions."
Real Samsung phone, Android 13 (720x1600, 300 dpi, en-GB), The Kana Quiz 20. License: MIT.
| App version | Android | Device | Skill version | Dates |
|---|---|---|---|---|
| 20 | 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.noprestige.kanaquiz (The Kana Quiz, GPL-3.0, version 20 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.
lblQuestion (the kana), lblResponse (prompt), lblScore ('Score: 50.0% / 1 / 2', only after the first answer) and one Button per choice (A I U E O for the first set).get_question()Launches the app and backs out of menus until lblQuestion shows.
answer(choice)Finds the button by text (case-insensitive). was_correct is inferred from the score counter rising.
get_score()Parses lblScore; 0/0 before the first answer.
The skill never knows the right answer: the agent reads the character in get_question and chooses. Question sets (hiragana, katakana, kanji) are chosen in the app's Question Selection menu, which this version does not drive.
Try a sample prompt or run an action by name. Replace any parameters in braces with your own values.
get_questionTry asking
ghost skill run ghost/kana-quiz get_question answerTry asking
ghost skill run ghost/kana-quiz answer get_scoreTry asking
ghost skill run ghost/kana-quiz get_score 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 | >=0.15 | pending | 14.5 KB | untagged |