luke: Skill for Claude Code

.agents/skills/changelog-screenshots/SKILL.md

changelog-screenshots is a skill for Claude Code, Codex from ReviewStage/luke. It costs 114 tokens per session (4,213 once invoked), scanned A, original, Apache-2.0.

A release-screenshot workflow for capturing pictures of features in a Mac build and adding them to a CHANGELOG.md file, which records software changes.

In plain words
What is it for?
It helps plan one purposeful image per relevant changelog section, stage mock data, capture the native app on macOS, and frame the images for the release.
Why use it?
It keeps screenshots tied to the exact code shipped in each release and prevents ad hoc pictures from missing important feature details.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

This is ReviewStage/luke's own configuration. It tells Claude Code and Codex how to work on luke itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything luke configures →

Reuse

Borrowing it

Nothing to install: this file belongs to ReviewStage/luke. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/ReviewStage/luke/main/.agents/skills/changelog-screenshots/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/ReviewStage/luke

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for changelog-screenshots

README.md
[![agentmods](https://agentmods.dev/badge/skills/reviewstage/luke/changelog-screenshots/github.svg)](https://agentmods.dev/skills/reviewstage/luke/changelog-screenshots)
Your own site
<a href="https://agentmods.dev/skills/reviewstage/luke/changelog-screenshots"><img src="https://agentmods.dev/badge/skills/reviewstage/luke/changelog-screenshots/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for changelog-screenshots

Your own site · 80×15
<a href="https://agentmods.dev/skills/reviewstage/luke/changelog-screenshots"><img src="https://agentmods.dev/badge/skills/reviewstage/luke/changelog-screenshots.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 114 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,213 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00114 $0.04213
Opus 5.5 $0.00046 $0.01685
Sonnet 5.5 $0.00023 $0.00843
Haiku 4.5 $0.00011 $0.00421

Measured 9d ago against content hash 9ca06c319fa4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

changelog-screenshots scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

The scan reads SKILL.md. This mod also ships 30 executable files (scripts/build.sh, scripts/frame-phone.py, scripts/patch-dist.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.agents/skills/changelog-screenshots/SKILL.md · 275 lines

How it starts

The opening of the file, as written. The whole thing — 275 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Changelog screenshots

The rules every picture follows are in the comment atop CHANGELOG.md; this is how to meet them. A release's pictures come from its own build: the commit the release's tag (v<version>, what shipped), or for a release not yet tagged, the head of its version pull request. The work is in four parts: plan the shots, stage them in a throwaway patch, build and capture on the Mac, then frame and land them.

You need the developer's Mac (a run-on-Mac tool, such as Luke's run_on_mac): the native build, the window capture, and the framer are macOS-only. The cloud workspace writes code and reviews pictures; with the workspace's file sync on, anything under its .context/ appears in the workspace's synced copy on the Mac ($S below), and nothing syncs back.

Without file sync (a Luke coding agent's sandbox), files cross through scratch refs on GitHub, never a command line, which cannot carry a patch of pictures: the cloud commits the staged worktree and pushes it to refs/luke-shots/<tag>, and the Mac fetches it into its clone and diffs it against the tag for build.sh; the Mac commits its previews and finished PNGs through a temporary index (GIT_INDEX_FILE, hash-object on absolute paths, write-tree, commit-tree) and pushes them to refs/luke-shots/out, which leaves the developer's checkout untouched. A ref outside refs/heads builds no Vercel Preview. Delete every scratch ref once the pictures land. The helpers run from a copy of origin/main's skill (git archive), not the developer's own checkout.

1. Plan the shots

Read the release's CHANGELOG.md entry and write a shot list in .context/shots/SPEC.md before any code: for each ### section, whether it gets a picture, what the picture must show, and the composition that puts that in front. Then look at the list as a set.

  • A section gets a picture only where one helps. Plumbing (a shared runtime, a fix) gets none; a single row or button is a close-up, not a whole window.
  • No two pictures share a composition. The 0.8.0–0.11.0 pictures fail this: seven of them are the same plan document on the left with the feature squeezed into a panel on the right. Vary the layout per shot: the sidebar folded or open, the side panel closed, wide, or full screen, a full-window view (Settings, version history), a menu, card, or palette held open.
  • Vary the framing: the whole window (1600 by 1000 points) where the layout is the point, a --crop close-up where the feature is one part of it. A close-up is the framed picture zoomed in, never a cut-out: the feature stays inside the app around it (the ⌘K palette over its dimmed plan, the sidebar with the window's corner and wallpaper beyond), and it comes out exactly as tall as a whole-window picture, only narrower. The changelog page shows every picture at that one height.
  • The plan document is not wallpaper. Show it only where it is the subject; elsewhere open another task, another tab, or fold it away.
  • Nothing experimental gets a picture. A feature behind Settings › Experimental (the Factory, Issues, Code review, video calls, and whatever joins them) is no section's subject, even where the notes announce it, and no shot turns an experiment on (ShotSpec.experiments stays empty), so nothing an experiment draws (a sidebar row, the Code review suggestion) shows at the edge of another picture either. 0.15.0's Factory and 0.17.0's Issues pictures were taken and then removed for this.

Read the full file on GitHub · 275 lines

Files

What ships with it

60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago Changed · +48 lines 9ca06c319fa4
  2. 16d ago Changed · -2 lines ff88785cd1c6
  3. 17d ago Changed · -1 lines 8f3bd70e3f9c
  4. 18d ago Changed · -1 lines 1563badb354e
  5. 19d ago Changed · +10 lines 9e03db86e0c3
  6. 21d ago Changed ed6d6adf3f75
  7. 22d ago Changed f844b956ec23
  8. 24d ago First seen · 221 lines · 114 tokens per session scan A 8856a4f65b82

Subscribe to this mod's changes

changelog-screenshots is a skill published in the GitHub repository ReviewStage/luke (39 stars, last pushed 5d ago), licensed Apache-2.0. It adds 114 tokens to every session and 4,213 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-14.

Related

changelog-screenshots compared with skills from other repositories

changelog-screenshots next to 6 skills from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
changelog-screenshotsReviewStage/lukeSkill395d agoApache-2.0A114Claude Code, Codex
imagegenopenai/codexSkill133kyesterdayApache-2.0A113Codex
playwright-climicrosoft/playwrightSkill99kyesterdayApache-2.0A19Claude Code
local-ai-agentsmicrosoft/ai-agents-for-beginnersSkill79k4d agoMITA200Claude Code, Codex
systematic-debuggingobra/superpowersSkill313k3d agoMITA21Claude Code
next-cache-components-adoptionvercel/next.jsSkill143kyesterdayMITA95Claude Code
next-partial-prefetching-adoptionvercel/next.jsSkill143kyesterdayMITA103Claude Code