aktual: Agent for Claude Code

Gradle Runner

Run a single Gradle command and return only the actionable output (errors, failures, warnings). This keeps the parent context window clean by absorbing verbose Gradle noise in an isolated agent.

Execution

Parse the caller's prompt for the Gradle task(s) to run and any extra flags, then run via the wrapper script which handles output capture, filtering, and exit-code propagation:

./scripts/gradle-run.sh <tasks> [flags]

Common invocations

# Compile one module (most common)
./scripts/gradle-run.sh :<module>:compileAll

# Run tests for one module - `:<module>:test` is AMBIGUOUS in this project; use a
# concrete task: `testAll`, `testAndroidHostTest`, or `testAndroid`
./scripts/gradle-run.sh :<module>:testAndroidHostTest --continue

# Format check
./scripts/ktfmt.sh check 2>&1 | grep -v "^$" | head -50

Getting more output

The script prints a filtered view, capped at 200 lines, and ends with the path of the full log, e.g. Full log (1234 lines): /.../build/gradle-run/20261006-191500-12345.log. If the filtered view is missing something the caller asked for, or says TRUNCATED, search that log instead of running Gradle again:

grep -nE "<pattern>" <log path> | head -100
sed -n '<from>,<to>p' <log path>

Test failure details are in <module-dir>/build/test-results/**/*.xml and detekt findings in <module-dir>/build/reports/detekt/issues.txt, with every module's findings from the last run merged in build/reports/detekt/all-issues.txt.

Helper scripts

./scripts/detekt.sh, compile.sh, test.sh and lint.sh aren't wrapped by gradle-run.sh. Run them directly and filter the output, e.g. ./scripts/detekt.sh 2>&1 | tail -100. Read detekt findings from build/reports/detekt/all-issues.txt and test failures from the result XML.

Output files

Never write output to a file of your own (> out.txt, tee, /tmp/...), not even a scratch one you delete afterwards. Everything you need is in the full log or under build/.