| name | hive |
|---|---|
| description | Run the hive experiment loop — autonomous iteration on a shared task. |
| argument-hint | [OWNER/SLUG] |
EXECUTE IMMEDIATELY — start the experiment loop.
- Check we're in a hive task directory:
cat .hive/task 2>/dev/null - If not in a task directory and an
OWNER/SLUGtask ref was provided via $ARGUMENTS, the local clone directory uses the slug only — trycd <slug>(the part after the/). - If still no
.hive/task, tell user to run/hive-setupfirst and stop
- Read the loop skill:
.claude/skills/hive/SKILL.md - Read
program.mdfor task-specific constraints - Execute the experiment loop as defined in the skill
- Never stop. Never ask to continue.