Development and testing

This page is for contributors making changes to auto-agents code or workflow assets.

Local development loop

Build quickly in debug mode.
Run targeted tests while iterating.
Run workspace-wide checks before opening a PR.

Build and run commands

# Build all crates
cargo build

# Run CLI help
cargo run -- --help

# Launch desktop app
cargo run -- gui

Test commands

# Full workspace tests
cargo test --workspace

# Targeted iteration example
cargo test agent::tracker::tests::

Quality gates

cargo fmt --all
cargo clippy --workspace --all-targets -- -D warnings

Treat clippy warnings as errors.

Hook and smoke utilities

# Install local hooks that mirror CI checks
./scripts/setup-hooks.sh

# Run smoke coverage for install/CLI surfaces
./scripts/smoke-test.sh

Test placement and naming

Prefer inline unit tests near the logic under test (#[cfg(test)]).
Use descriptive snake_case test names.
Add coverage for parser changes, workflow prompt generation, and command construction paths.

Commit and PR expectations

Keep commits focused and imperative.
In PR descriptions, summarize user-visible change and list verification commands run.
Include screenshots or short recordings when changing UI behavior.

Documentation update policy

When changing workflows, command behavior, or adapter semantics, update both:

Project README/docs where needed.
Writerside topics in this directory for structured product documentation.

16 July 2026