mjolnir

Mjölnir entry — create a new branch in the current working tree, then route to the appropriate skill.

Mjölnir entry — proceed on the current branch and worktree, then route to the appropriate skill.

Mjölnir entry — create a worktree for this work, then route to the appropriate skill.

Use this agent when a major project step has been completed and needs to be reviewed against the original plan and coding standards. Examples: <example>Context: The user is creating a code-review agent that should be called after a logical chunk of code is written. user: "I've finished implementing the user authentication system as outlined in step 3 of our plan" assistant: "Great work! Now let me use the code-reviewer agent to review the implementation against our plan and coding standards" <commentary>Since a major project step has been completed, use the code-reviewer agent to validate the work against the plan and identify any issues.</commentary></example> <example>Context: User has completed a significant feature implementation. user: "The API endpoints for the task management system are now complete - that covers step 2 from our architecture document" assistant: "Excellent! Let me have the code-reviewer agent examine this implementation to ensure it aligns with our plan and follows best practices" <commentary>A numbered step from the planning document has been completed, so the code-reviewer agent should review the work.</commentary></example>

Use when invoked by a Mjölnir entry command or when the human partner asks for help designing a feature, refining requirements, or exploring approaches before any implementation. Surfaces intent and trade-offs through dialogue and produces a written spec at <operating-root>/.mjolnir/specs/.

Use when invoked by a Mjölnir entry command or when the human partner has 2+ independent tasks that can run without shared state. Fans them out to subagents with isolated context, then aggregates results. Best for parallel research, multi-area exploration, or independent fixes.

Use when the human partner is responding to code review feedback (from mjolnir:code-reviewer or from a human reviewer). Encourages technical rigor — restate the requirement, verify the claim, evaluate, then implement one item at a time. Not performative agreement, not blind compliance.

Use when invoked by a Mjölnir entry command or when the human partner has completed a meaningful chunk of work and wants it reviewed. Dispatches the mjolnir:code-reviewer subagent with a crafted context (never session history) so the reviewer focuses on the work product, not your thinking trail.

Use when invoked by a Mjölnir entry command (/mjolnir:worktree, /mjolnir:branch, or /mjolnir:current) to dispatch the work to the appropriate downstream skill based on what the human partner described.