What kind of task should I choose first?
Choose a low-risk feature or cleanup with clear behavior, no sensitive data migration, and a review scope you can inspect in one sitting.
Docs
Run a small Hal loop from PRD planning through validation, implementation, and human review without handing over a broad feature.
Why it matters
The safest first Hal loop is intentionally small: one bounded requirement, clear acceptance criteria, and a developer review before merge.
Run a first Hal loop with one small requirement, a clean branch, validated stories, and a manual review before merge. The goal is not hands-off shipping. The goal is to learn the PRD-native sequence: initialize, plan, convert, validate, run, inspect, and decide what to keep.
Create a branch for the experiment before the agent edits files:
git checkout -b hal-first-loop
A clean branch makes every generated change easier to review or discard.
Run the project setup command:
hal init
This creates the local workflow structure Hal uses for planning, validation, runtime state, and reviewable artifacts.
Use a narrow feature that has a visible outcome:
hal plan "add a dark mode toggle with system preference detection"
Review the generated PRD or story output before moving forward. Look for ambiguous scope, missing constraints, and unclear acceptance criteria.
Prepare runtime state and check story quality:
hal convert
hal validate
If validation reports unclear stories, fix the PRD before implementation. A vague story usually becomes a vague diff.
When the plan is reviewable, run the loop:
hal run
Hal can run autonomous implementation iterations with the configured engine, but it should still leave state a developer can inspect.
Inspect changed files, commits, reports, tests, generated state, and remaining acceptance criteria. Revert or repair anything that does not match the PRD.
Use Hal to make agent work more structured, not to remove developer responsibility.
Before adopting Hal, verify the current source repository, install path, release notes, and supported engine behavior. A completed loop is reviewable agent output, not a guarantee that the code is safe to merge.
Use the PRD readiness checker before planning, read the PRD-driven planning feature, and review whether Hal replaces code review before expanding to larger tasks.
Short answers before you put this into an agent workflow.
Choose a low-risk feature or cleanup with clear behavior, no sensitive data migration, and a review scope you can inspect in one sitting.
Hal is designed for PRD-native loops. Use planning and validation so the implementation engine has a bounded target.
No. Treat the result as reviewable agent output. Inspect files, commits, tests, reports, and unresolved acceptance criteria before merging.
Keep exploring the pieces of a reviewable coding loop.
Install the Hal CLI, verify the command is available, and check the AI coding engine prerequisites before running a PRD-native loop.
Use this checklist to review AI coding loop output from Hal before merging commits, reports, workflow state, or generated code.