argos-pr-review

Reviews Argos visual regression builds and submits anchored feedback during pull request review.

Updated Sep 15, 2026
One-click install
npx skills add https://github.com/Nadav011/nadavai-claude --skill argos-pr-review-nadav011
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: argos-pr-review
Source: https://github.com/Nadav011/nadavai-claude/tree/main/plugins/nadavai/skills/argos-pr-review
Command: npx skills add https://github.com/Nadav011/nadavai-claude --skill argos-pr-review-nadav011

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When a pull request triggers an Argos visual regression build with detected changes, someone must decide whether each screenshot diff is intentional, a regression, or a flaky capture before approving. This Skill turns that manual screenshot-by-screenshot judgment into a structured, evidence-based review workflow. ## Core Features & Use Cases - Build status triage: Inspects the Argos build via the argos CLI and routes by status (changes-detected, pending, rejected, orphan) before doing any review work. - Flakiness analysis: Weighs test history metrics (flakiness, stability, occurrences) against capture-level signals (spinners, retries, mid-animation states) to distinguish real regressions from noise. - Anchored review output: Posts per-diff comments pinned to regions, submits approve/reject reviews, and can silence confirmed recurring flakes with change ignore. - Use Case: A PR shows an Argos status check with 12 changed snapshots. The Skill fetches the diffs, identifies 10 as intentional style updates, flags 1 clipped button as a regression with an anchored comment, and marks 1 recurring loader flake as ignorable, then submits the review. ## Quick Start Review the Argos build linked in this pull request and tell me whether the visual diffs match the intended changes.

Frequently Asked Questions about argos-pr-review

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

FAQPage Schema
How do I review an Argos visual regression build in a pull request?

Fetch the build with argos build get, then list changed snapshots with argos build snapshots --needs-review --json. Judge each diff against the PR's intended change, post anchored comments on problem diffs, and submit an approve or reject review with argos review create.

How to tell if an Argos screenshot diff is flaky or a real regression?

Combine two signals: test history metrics like flakiness score, change occurrences, and stability, plus capture-level signs like visible loaders, mid-animation states, or retry counts. When both agree the diff is environmental noise, classify it as flaky; a stable test that changed is more likely intentional or a real regression.

What tokens does the Argos CLI need for build review?

Reading builds and snapshots needs a project token via ARGOS_TOKEN or --token. Submitting reviews or comments requires a personal access token from argos login. Without a PAT, you can still report conclusions but cannot post the review.

Can I ignore a recurring flaky diff so it stops blocking builds?

Yes, use argos change ignore with the change ID and a personal access token to silence a confirmed recurring flake, and reverse it with change unignore. Only ignore flakes the metrics confirm; prefer a code fix like aria-busy or data-visual-test attributes when one exists.

What does an Argos orphan build mean during review?

An orphan build has no prior baseline to compare against, so all snapshots appear as added with a null base. This is expected for first builds or new branches and is not itself a regression signal; approve it if screenshots look correct to seed a future baseline.