review-queue

Triages blocked auto-dev pipeline items and executes human decisions.

515|64|Updated Oct 23, 2024
One-click install
npx skills add https://github.com/allenhutchison/obsidian-gemini --skill review-queue
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-queue
Source: https://github.com/allenhutchison/obsidian-gemini/tree/main/.agents/skills/review-queue
Command: npx skills add https://github.com/allenhutchison/obsidian-gemini --skill review-queue

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The maintainer's daily console for the auto-dev pipeline surfaces everything currently blocked on you (the open automated PR awaiting review/merge, plans awaiting approval, questions and park proposals awaiting an answer, parked issues to revisit, and brand-new untriaged issues), presents it as a one-line worklist, then loops: you reply with an item number and your decision, the skill executes that decision against the auto:* state machine as you (never with the auto-dev marker), and re-presents the shrinking list until the queue is drained or you say you're done. Use when the user says "review the pipeline", "what needs my input", "review the issue queue", "process the auto-dev queue", "check what auto-dev is waiting on", "let's do the daily review", "approve/answer/park #N", or similar. Never merges on its own. Never builds issues — it only feeds your decisions back to the pipeline.

Core Features & Use Cases

  • The human half of the auto-dev pipeline surfaces blocked items and requires your decisions to progress.
  • Executes decisions against the auto:* state machine while ensuring human oversight and prohibiting automatic merges.
  • Maintains safety invariants: you are the human, never auto-merge, and decisions are posted as your input.

Quick Start

Start by listing the current queue, then reply with decisions for each item (e.g., '#663 approved' or 'park #641'), and the skill will apply your choices and refresh the worklist.

Frequently Asked Questions about review-queue

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

FAQPage Schema
How do I triage blocked pull requests and issues in an automated GitHub workflow?

To review blocked auto-dev pipeline items, this skill gathers pending pull requests, untriaged issues, and plans awaiting approval, presents them as a one-line worklist, and loops your decisions through the auto:* state machine without auto-merging.

What is a human-in-the-loop review queue for GitHub pull requests?

A human-in-the-loop review queue for GitHub pull requests is a daily console that surfaces automated PRs awaiting review, presents them for explicit approval, and enforces human responsibility by prohibiting automatic merges throughout the session.

How do I process the auto-dev queue and apply decisions to GitHub issues?

You process the auto-dev queue by listing blocked items, replying with decisions like '#663 approved' or 'park #641', and the skill applies your choices to GitHub issues via the auto:* state machine, then refreshes the worklist until drained.

Can I use this review queue to automatically merge pull requests on GitHub?

No, you cannot automatically merge pull requests on GitHub with this review queue. It enforces safety invariants that prohibit automatic merges, requiring your explicit approval before any merge decision is executed against the pipeline.

Does the review queue work with untriaged GitHub issues and parked proposals?

Yes, the review queue works with untriaged GitHub issues and parked proposals by gathering everything currently blocked on you, including questions and park proposals awaiting an answer, and presenting them for your decision.

What are the limitations of using a human-in-the-loop triage workflow for auto-dev pipelines?

The limitations of a human-in-the-loop triage workflow for auto-dev pipelines include requiring explicit human input for every decision, never building issues independently, and only feeding your decisions back to the pipeline without introducing auto-dev automation.