skills.rest
kimprobably
kimprobablyCOMMUNITY
Explore all skills by this author →
GitHub

spec-kitty-implement-review

Orchestrate Spec Kitty work package implementation and review loops with dependency-aware sequencing.

Updated May 13, 2026
Search Tags:#agent dispatch#code review automation#dependency sequencing#work package orchestration#spec kitty#rejection loop#arbiter escalation
One-click install
npx skills add https://github.com/kimprobably/maestro-os --skill spec-kitty-implement-review
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: spec-kitty-implement-review
Source: https://github.com/kimprobably/maestro-os/tree/main/.claude/skills/spec-kitty-implement-review
Command: npx skills add https://github.com/kimprobably/maestro-os --skill spec-kitty-implement-review

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill automates the end-to-end implement-and-review workflow for Spec Kitty work packages, including dependency-aware sequencing and robust handling of review rejections.

Core Features & Use Cases

  • Dependency-aware WP sequencing: Uses the dependency graph to decide what to run next and when dependents can start.
  • Two-step agent workflow: Separates implementation dispatch from review dispatch using configured agents.
  • Rejection-cycle management: Tracks review rejection cycles (max 3), then switches to arbiter mode for escalation or approval decisions.
  • Parallel sprint execution: Supports claiming multiple WP workspaces for parallel agent execution while preserving lane constraints.

Quick Start

Orchestrate a full implement-review sprint by asking the system to run the implement-review loop for all WPs using the WP scheduler input: spec-kitty-implement-review all

Frequently Asked Questions about spec-kitty-implement-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code review workflows with dependency-aware task sequencing?▼

Automate code review workflows by orchestrating implementation and review dispatch across coding agents using a dependency graph. This dependency-aware sequencing decides execution order and safely starts dependent tasks, tracking state transitions from planned to approved.

What is a rejection loop in automated code review and how does escalation work?▼

A rejection loop in automated code review tracks iterative review failures up to a maximum of three cycles. After repeated rejections, the system switches to arbiter mode for escalation, managing final approval decisions or manual handling for unstable agents.

How do I run parallel agent execution for multiple work packages?▼

Run parallel agent execution by claiming multiple work package workspaces simultaneously while preserving lane constraints. This work package orchestration supports parallel sprint execution alongside the two-step implementation and review agent dispatch workflow.

Can I use manual handling for coding agents lacking stable CLI capabilities?▼

Manual handling is supported for coding agents lacking stable CLI capabilities during work package orchestration. Mandatory coordination of agent dispatch and state transitions still applies, ensuring the implement and review loop completes properly.

What are the limitations of automated rejection cycle management for code review?▼

Automated rejection cycle management is limited to a maximum of three review rejection cycles. Once this limit is reached, the system stops iterating and escalates to arbiter mode for final resolution or manual intervention.

Related Skills

Explore More Software Engineering Skills→
koolamusickoolamusic
9

spawn

Decompose software plans into parallel agent tasks with worktree isolation.

Community
Advanced
danilonovaisvdanilonovaisv
1

subagent-driven-development

Dispatch isolated subagents per task and run spec-compliance and code-quality reviews.

Community
Advanced
liauw-medialiauw-media
3

dispatching-parallel-agents

Dispatch parallel agents to execute independent tasks concurrently.

Community
Advanced
tonyandrewmeyertonyandrewmeyer
1

iterate-fix

Triage build, deploy, and test failures with bounded retries and escalation.

Community
Advanced