heruujoko/mothership

mothership

Use when the user explicitly invokes /mothership or asks to run the mothership orchestration workflow manually in a repository that may also have many other installed skills.

model-policy-setup

Bootstrap mothership model policy config. Creates mothership-config/model-policy.yaml if missing, or overwrite with --force.

mothership-setup

Configure the mothership wiki for the current project. Run once per project to set up wiki directory structure, preferences, and default schema. Re-run to change wiki root location or other settings.

pr-maintainer

Maintain a GitHub pull request through the GitHub CLI by creating PRs, reading review feedback, routing each feedback item back through mothership, and updating the PR until it is ready for human-approved merge.

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

creating-plan

Bundled dependency skill used by mothership research for structured plan creation.

executing-plans

Bundled dependency skill used by mothership coding phase to execute an approved implementation plan step-by-step without skipping back into ad hoc design.

making-plans

Bundled dependency skill used by mothership intake, research, and planning phases to turn explored requirements into an explicit implementation plan before coding starts.

superpowers

Bundled dependency skill used by mothership research for broad context gathering and option mapping.

research-execution

Define how research agents investigate a task before implementation planning. Covers hypothesis formation, context gathering, and structured output.

github-issue-management

Define how mothership uses GitHub issues as durable workflow artifacts. Covers issue creation, linking, status tracking, and closure hygiene.

human-escalation

Define when and how the commander requests human input during the mothership workflow. Covers escalation triggers and escalation handling.

systematical-debugging

Bundled dependency skill used by mothership coding after QA feedback for structured debugging.

parallelization-decision

Help commander decide whether work should be assigned to one coder agent or multiple coder agents in parallel.

qa-review-loop

Define the review cycle between coder agents, QA, and commander. Covers feedback routing, iteration, and approval criteria.

risk-assessment

Give commander a structured way to classify task risk before choosing execution strategy. Covers risk dimensions and mitigation planning.

task-intake-and-decomposition

Help commander turn an incoming human request into an actionable execution plan. Covers scoping, decomposition, and success criteria definition.

worktree-management

Define when and how isolated git worktrees are used for coder agents. Covers worktree creation, allocation, and cleanup.

commit-push

Commit scoped local changes and push the current branch to its remote without opening a pull request. Use when the user explicitly asks to commit and push, publish a branch, or update an existing remote branch but does not ask to create a PR.

commit

Create a local git commit for a scoped set of changes without pushing. Use when the user asks to commit current work, make a commit message, or checkpoint changes locally but does not ask to push or open a pull request.

create-pr

Create a pull request for the current branch with a real title and Markdown description after the branch exists on the remote. Use when the user asks to open a PR, draft a PR, or create a pull request description from the current changes.