메뉴
Create or resume a treehouse-managed leased worktree for one feature or subfeature, drive `ship` inside that worktree, and hand the finished branch back as a pull request to the original branch.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Bootstraps `.skills/planning.json`, `.skills/execution.json`, and `.skills/conventions.json` for a repository, and can also scaffold a wiki knowledge layer next to the planning feature directory. Use when a user asks to configure a project, initialize `sirius-skills` settings, apply generic defaults, add Jira-oriented conventions, or bootstrap the repo's wiki skeleton.
Canonical workflow for creating and managing durable subfeatures under an existing feature. Use this whenever the user asks to add, create, start, open, or bootstrap a subfeature so downstream planning skills keep the expected metadata and discovery-draft lifecycle.
Reconcile one planning target, optionally execute planning owners in sequence, and persist checkpointed resume context through approval, optional auto-commit handoff, and planning-commit handoff.
Converts repo stories and planning docs into directly executable, dependency-aware work items with traceability and execution handoff.
Closes the active execution slice with validation and records durable closure metadata while invoking shared owner-completion reconciliation.
Produces feature-level system-design.md artifacts before breakdown when a feature needs architecture, interface, constraint, failure-handling, or validation decisions captured durably.
| name | ship-worktree |
| description | Create or resume a treehouse-managed leased worktree for one feature or subfeature, drive `ship` inside that worktree, and hand the finished branch back as a pull request to the original branch. |
Use this skill when a feature or subfeature should be implemented in its own treehouse-managed worktree instead of on the current branch.
ship
backlog surface.ship --resume or ship --finalize inside the leased worktree without
moving backlog ownership out of ship and ship-slice.sirius ship-worktree <target> --json
sirius ship-worktree <target> --resume --json
sirius ship-worktree <target> --finalize --json
sirius ship-worktree <target> --create-pr --json
sirius ship-worktree <target> --resume --create-pr --json
treehouse get --lease --lease-holder <target>
treehouse return <worktree-path> --force
Treehouse owns the worktree pool configuration.
# Repo-level treehouse config
max_trees = 16
root = "$HOME/.treehouse"
[hooks]
post_create = ["./scripts/setup-venv.sh"]
pre_destroy = ["./scripts/teardown.sh"]
Use treehouse init to write the repo config, and use
~/.config/treehouse/config.toml for user-level settings.
ship_worktree.py <target> --json to resolve the target and ensure the
leased worktree exists.--resume to continue the feature or subfeature backlog from inside that
leased worktree through the existing ship orchestration.--finalize when all slices are completed and the target is ready for
final archive/reconciliation handling.--create-pr when the worktree branch is ready to open review back to
the original base branch.ship, ship-slice,
review-execution, close-slice, reconcile-execution, and commit.