ghost/calculator

Fossify Calculator: evaluate arithmetic, read the display and history, convert units. No account, no network.

Proven 12.3× fewer tokens Score 39/100 Official hard skill python

Before / after

Benchmark details ↓
Tokens saved123,150Paired medians · 1/1 tasks · N=5
UI actions saved4Paired medians · 1/1 tasks · N=5
Pass · before → with skill5/5 → 5/5Current arms · failures included

Before With skill · median tokens per task

Calculator "3+3" Calculator "3+3". Median tokens: before 184,681, with skill 61,837. N=5 before / 5 with skill. 184,68161,837 −67% tokensN=5/5 N=5 before / 5 with skill
Before 5/5 · Wilson 95% CI 0.566–1.000; with skill 5/5 · Wilson 95% CI 0.566–1.000

3 more actions not benchmarked yet

maintainer-run, not independent · claude-haiku-5-5 · 2026-10-22. Every failure is counted (intention-to-treat).

Demo: calculate · HOME → launch → action

Control Calculator on Android with an AI agent

Demo shows: convert_units (category=Temperature, value=100, from_unit=Fahrenheit, to_unit=Celsius)

Ghost in the Droid seed skill for Fossify Calculator (org.fossify.math), a FOSS calculator from F-Droid. Offline, no account.

ActionWhat it doesParams
calculateKey in an expression, press =, return the resultexpression
read_resultReturn the number on the displaynone
get_historyReturn recent calculations from the History dialognone
convert_unitsConvert a value with the built-in unit convertercategory, value, from_unit, to_unit

Note: the app evaluates left to right (84-2*6 gives 492).

Install

The app: F-Droid, or https://f-droid.org/packages/org.fossify.math/. 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/calculator

Use

ghost skill run ghost/calculator calculate --param expression="1234*5" --device emulator-5554
ghost skill run ghost/calculator convert_units --param category=Temperature --param value=100 \
  --param from_unit=Fahrenheit --param to_unit=Celsius --device emulator-5554
run_skill(name="ghost/calculator", action="calculate", params={"expression": "100/8"})
# {"display": "12.5", "value": 12.5}

Tested on

Android 14 emulator (google_apis x86_64, 1080x2400, 420 dpi), Fossify Calculator 1.4.0, en-US. Signed self-runs in runs/. License: MIT.

Known failure modes and versions tested

Versions tested

App versionAndroidDeviceSkill versionSigned self-runs (pass/total)Dates
1.4.014Google/sdk_gphone64_x86_64 (emulator, 1080x2400, 420 dpi, en-US)1.0.014/142026-10-05

Android 13 was additionally replayed on a 720x1600 Samsung SM-A037F (300 dpi, en-GB) and a 1080x2400 emulator (420 dpi, en-US). See the cross-device report for measured outcomes and limitations; other profiles remain unverified.

Historical self-run artifacts are retained in runs/, but the current registry rejects the pre-digest evidence as unbound-helper. These old counts do not establish current-version passing coverage; fresh digest-bound signed runs are pending.

Recorded failures

The historical runs above were green. Samsung failures and subsequent fixes are recorded in the cross-device report.

Popups and first-run screens handled

  • None app-specific; Android permission dialogs go through the engine's OS popup layer.

Other known failure modes

  • A unit name that the app does not list fails with the list of offered units in the error.
  • The keypad loop rejects unmapped characters after launch and clear, and can have tapped a prefix already. The validator accepts numeric expressions with + - * / ^ and multiplication/division aliases x, ×, ÷; % has a keypad mapping but is not accepted by the arithmetic validator.
  • Arithmetic validation evaluates only the requested operator; zero right operands work for addition, subtraction and multiplication. Actual division by zero fails.

Compatibility and validation

The manifest targets Android >=13 and Fossify Calculator >=1.4.0, with en-US and en-GB flows. Its authoring screen is 1080x2400 at 420 dpi; measured Samsung/emulator outcomes are in the cross-device report. calculate rereads the device display and compares the left-to-right result. read_result independently rereads the display; get_history leaves History open and rereads its visible rows; convert_units rereads input, output, both unit names and the symbol, comparing each with the returned data. Conversion also checks caller input/unit names; it does not independently recompute conversion mathematics.

Try a sample prompt or run an action by name. Replace any parameters in braces with your own values.

calculate

Play demo

Try asking

  • Calculate 1234 times 5
  • calculate 1234 times 5
  • use the calculator to divide 100 by 8
  • what's 2 to the power of 10
ghost skill run ghost/calculator calculate

read_result

Play demo

Try asking

  • What does the calculator show right now?
  • read the calculator display
  • what's the current result on the calculator
ghost skill run ghost/calculator read_result

get_history

Play demo

Try asking

  • Show my recent calculations
  • calculator history
  • what did I calculate before
ghost skill run ghost/calculator get_history

convert_units

Play demo

Try asking

  • Convert 100 Fahrenheit to Celsius
  • how many kilometers is 26.2 miles
  • convert 70 kilograms to pounds
  • what is -40 celsius in fahrenheit
ghost skill run ghost/calculator convert_units
Benchmark details

Ghost-native before/after

maintainer-run, not independent
Model: claude-haiku-5-5 · Date: 2026-10-22

Same task across three arms. Skill replay runs without a model. Medians include every valid episode, including failures.

Calculator "3+3"

N=5 seeds · 2026-10-22

Arm / provenancePass a/b · Wilson 95% CITokens, median [q1–q3]LLM calls, medianUI actions, medianWall, median
Agent aloneclaude-haiku-5-5 · 2026-10-22
Skill 1.0.2 · SHA b37d4c36
5/5 · Wilson 95% CI 0.566–1.000184,681[132,498–267,125]6642 s
Skill replayno model · 2026-10-22
Skill 1.0.2 · SHA b37d4c36
5/5 · Wilson 95% CI 0.566–1.0000[0–0]0643 s
Agent + skillclaude-haiku-5-5 · 2026-10-22
Skill 1.0.2 · SHA b37d4c36
5/5 · Wilson 95% CI 0.566–1.00061,837[61,677–111,942]3160 s

Agent + skill vs agent alone: 5 paired seeds; 0 fail → pass, 0 pass → fail; McNemar exact p = 1. Median tokens saved: 123,150.

Skill replay vs agent alone: 5 paired seeds; 0 fail → pass, 0 pass → fail; McNemar exact p = 1. Median tokens saved: 184,681.

Episodes (15)
TaskArmSeedModelPassLLM callsTokensUI actionsWallTrace
Calculator "3+3"Agent + skill0claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass362k160 s
Calculator "3+3"Skill replay0replay2026-10-22 · Skill 1.0.2 · SHA b37d4c36pass00643 s
Calculator "3+3"Agent alone0claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass8185k655 s
Calculator "3+3"Agent + skill1claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass6254k249 s
Calculator "3+3"Skill replay1replay2026-10-22 · Skill 1.0.2 · SHA b37d4c36pass00641 s
Calculator "3+3"Agent alone1claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass6132k646 s
Calculator "3+3"Agent + skill2claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass362k162 s
Calculator "3+3"Skill replay2replay2026-10-22 · Skill 1.0.2 · SHA b37d4c36pass00643 s
Calculator "3+3"Agent alone2claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass6115k642 s
Calculator "3+3"Agent + skill3claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass362k160 s
Calculator "3+3"Skill replay3replay2026-10-22 · Skill 1.0.2 · SHA b37d4c36pass00641 s
Calculator "3+3"Agent alone3claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass8288k537 s
Calculator "3+3"Agent + skill4claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass4112k167 s
Calculator "3+3"Skill replay4replay2026-10-22 · Skill 1.0.2 · SHA b37d4c36pass00643 s
Calculator "3+3"Agent alone4claude-haiku-5-52026-10-22 · Skill 1.0.2 · SHA b37d4c36pass6267k537 s

Display only: points come from signed run artifacts.