skills.rest
Priivacy-ai
Priivacy-aiOFFICIAL
Explore all skills by this author →
GitHub

spk-run-implement-review

Orchestrate Spec Kitty work-package implementation and review loops until completion.

1.6k|156|Updated Oct 9, 2025
Search Tags:# orchestration#work packages#Code review#spec-Driven Development#Spec Kitty#AI Agents#Git Worktrees
One-click install
npx skills add https://github.com/Priivacy-ai/spec-kitty --skill spk-run-implement-review
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: spk-run-implement-review
Source: https://github.com/Priivacy-ai/spec-kitty/tree/main/src/charter/offering/skills/spk-run-implement-review
Command: npx skills add https://github.com/Priivacy-ai/spec-kitty --skill spk-run-implement-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinating multiple AI agents through implementation and review cycles is error-prone: work packages get stuck in lanes, reviewer feedback is lost, and sub-agents appear hung while synchronous pre-review test gates are still running. This Skill drives the full implement-review loop for a Spec Kitty mission until every work package reaches done, approved, or a correctly rejected state.

Core Features & Use Cases

  • Loop Orchestration: Identifies the active work package and lane via spk-run-next, dispatches implementation from runtime prompt context, and reviews each package independently with spk-run-review-wp.
  • Rejection Feedback Loop: Feeds structured reviewer feedback back into the next implementation attempt so rejected work packages converge on acceptance.
  • Long-Gate Contract Handling: Treats move-task --to for_review as a synchronous pre-review regression gate that must be polled to completion, with opt-outs via --skip-pre-review-gate or the SPEC_KITTY_SYNC_DISABLE environment variables.
  • Use Case: A team running parallel Claude Code or Codex agents in git worktrees uses this Skill to keep implementers and reviewers cycling through work packages without manual lane management.

Quick Start

Ask the agent to run the implement-review loop for the current Spec Kitty mission and keep cycling work packages until all lanes reach done or approved.

Frequently Asked Questions about spk-run-implement-review

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

FAQPage Schema
How do I run an implement and review loop for Spec Kitty work packages?▼

Use spk-run-next to identify the active work package and lane, dispatch implementation from the runtime prompt context, then review each package with spk-run-review-wp. Repeat the cycle, feeding reviewer feedback into retries, until the runtime returns a terminal state.

Why does my sub-agent hang on move-task to for_review?▼

The for_review transition runs a synchronous pre-review regression gate that executes a scoped test subprocess before returning, which can take seconds to minutes. Poll the command to completion instead of killing or retrying it while the gate is still running.

How do I skip the pre-review regression gate in Spec Kitty?▼

Pass --skip-pre-review-gate to skip the gate for a single move-task invocation. To disable it process-wide, set the SPEC_KITTY_SYNC_DISABLE or SPEC_KITTY_SYNC_MINIMAL_IMPORT environment variables before running the command.

What happens when a work package review is rejected?▼

The structured reviewer feedback is fed back into the next implementation attempt for that work package. The loop continues dispatching implementation and review cycles until the package is approved, done, or the runtime reports a blocker.

When should I use spec-kitty-implement-review instead?▼

Use the spec-kitty-implement-review legacy alias when you need detailed orchestration mechanics beyond the high-level loop contract. The spk-run-implement-review skill covers the standard flow and the sub-agent long-gate behavior.

Related Skills

Explore More Software Engineering Skills→
EtroxTaranEtroxTaran

plan

Generate task breakdowns with acceptance criteria, file scope, and test plans.

Community
Basic
shuff57shuff57

parallel-build

Orchestrate parallel git worktree builds with human-gated merges.

Community
Advanced
guifryguifry

ralph-status

Assemble Ralph loop status reports from PRD.json, progress.txt, and git log.

Community
Intermediate
adawadaw

fabric-hotfix

Automate urgent XS/S fixes through analysis, coding, testing, review, and merging.

Community
Advanced