open_issues
Play demoTry asking
- Open the Decisions pros and cons app
- open Decisions
ghost skill run ghost/decisions open_issues Decisions (F-Droid): weigh a decision on a balance. Create an issue, add pros and cons with a weight from -10 to 10 and watch the scale tip. 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.
17.3× fewer tokens, 2.1× 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 | 181 s | 1,562,788 | 68 | stopped early |
| With skill: find_skill, run_skill | 85.7 s | 90,508 | 7 | finished |
Prompt: Start a pros and cons list called Buy a new phone, add the pro Cheaper with weight 5 and the con Worse camera with weight -4, then read back the list. 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 Decisions on Android with an AI agent. Can't decide? Your agent lists the pros and cons with weights and the scale shows the verdict.
Target search phrase: AI agent weighs pros and cons decision on Android
Demo shows: new_issue (title=Buy a new phone)
Ghost in the Droid seed skill for Decisions (
de.markusfisch.android.libra), an offline pros and cons tool with a balance that tips as you add weighted arguments. Titles and argument chips are text nodes, so an agent can weigh a decision. Offline, no account.
| Action | What it does | Params |
|---|---|---|
open_issues | Restart the app and read the list of saved issues. | none |
new_issue | Create a new issue with a title, ready for arguments. | title |
add_argument | Add an argument with a weight from -10 (against) to 10 (for) to the open issue and watch the scale tip. | text, weight |
get_issue | Read the open issue: title and every argument with its side (pro, con or neutral). | none |
list_issues | Go back to the issue list and read every saved issue with its age. | none |
An agent can drive the whole app with these; compose them.
| Role | Action |
|---|---|
| Entry | open_issues, new_issue |
| Primitives | add_argument |
| Read state | get_issue, list_issues |
The app: F-Droid, or https://f-droid.org/packages/de.markusfisch.android.libra/. 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/decisions
ghost skill run ghost/decisions add_argument --param "text=Cheaper" --param weight=5 --device <serial>
run_skill(name="ghost/decisions", action="add_argument", params={"text": "Cheaper", "weight": 5})
# {"result": {"text": "Cheaper", "weight": 5, "side": "pro"}}
Try asking: "Start a pros and cons list called Buy a new phone, add the pro Cheaper with weight 5 and the con Worse camera with weight -4, then read back the list."
Real Samsung phone, Android 13 (720x1600, 300 dpi, en-GB), Decisions 25. License: MIT.
| App version | Android | Device | Skill version | Dates |
|---|---|---|---|---|
| 25 | 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: de.markusfisch.android.libra (Decisions, MIT, version 25 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.
open_issues()Force-stops and relaunches.
new_issue(title)Taps add, renames the issue through the pencil dialog and confirms.
add_argument(text, weight)Types the argument, taps the weight number on the scale row and saves.
get_issue()Reads the title and the argument chips; the side comes from the chip position.
list_issues()Navigates up from the issue and reads the list rows.
The weight number and the balance totals are drawn on a canvas, not in the accessibility tree: the skill reads a chip's side (pro, con, neutral) from its position and echoes the weight it set, but cannot read the weight back. Pass weight 0 for a neutral argument. Texts are typed through adb, so keep them plain.
Try a sample prompt or run an action by name. Replace any parameters in braces with your own values.
open_issuesTry asking
ghost skill run ghost/decisions open_issues new_issueTry asking
ghost skill run ghost/decisions new_issue add_argumentTry asking
ghost skill run ghost/decisions add_argument get_issueTry asking
ghost skill run ghost/decisions get_issue list_issuesTry asking
ghost skill run ghost/decisions list_issues 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 | 15.7 KB | untagged |