Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add natsukium/dotfiles --skill gui-debuggit clone --depth 1 https://github.com/natsukium/dotfilesWrote 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.
[](https://agentmods.dev/skills/natsukium/dotfiles/gui-debug)<a href="https://agentmods.dev/skills/natsukium/dotfiles/gui-debug"><img src="https://agentmods.dev/badge/skills/natsukium/dotfiles/gui-debug/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.
<a href="https://agentmods.dev/skills/natsukium/dotfiles/gui-debug"><img src="https://agentmods.dev/badge/skills/natsukium/dotfiles/gui-debug.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00089 | $0.00986 |
| Opus 5.5 | $0.00036 | $0.00394 |
| Sonnet 5.5 | $0.00018 | $0.00197 |
| Haiku 4.5 | $0.00009 | $0.00099 |
Grade A, and why
gui-debug scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
macOS GUI debugging
Verify a rendering or window-chrome change by launching the app,
capturing just its window, and probing pixels. None of this needs
Accessibility permission. (System Events / keystroke injection does
need it and is usually blocked — don't plan on typing into the window
unattended.) Screen Recording permission is required for screenshots:
if screencapture -x /tmp/full.png produces a non-black image, you
have it.
Throughout, replace MyApp with the app's window-owner name (what
shows in the menu bar / Activity Monitor) and myapp with the process
name for pkill -x.
Launch → window id → capture
./path/to/MyApp 2>/tmp/app.log & sleep 4 # or: open -a MyApp
WID=$(osascript -l JavaScript -e '
ObjC.import("CoreGraphics");
const list = ObjC.castRefToObject($.CGWindowListCopyWindowInfo($.kCGWindowListOptionOnScreenOnly, $.kCGNullWindowID));
const arr = ObjC.deepUnwrap(list);
const w = arr.filter(x => x.kCGWindowOwnerName === "MyApp")[0];
w ? w.kCGWindowNumber : "none";')
screencapture -o -l$WID /tmp/app-win.png # window only, no drop shadow
ObjC.castRefToObjectis required —ObjC.deepUnwrapon the rawCFArrayReffails with "Ref has incompatible type".screencapture -o -l<id>captures the window's own composited content with per-pixel alpha: a translucent background shows its real alpha (e.g. 153 foropacity = 0.6), and rounded-corner pixels read fully transparent. This is the ground truth for transparency / corner-clip checks — a full-screen capture is not.- A tiling WM (or the app itself) may resize the window after launch —
re-read
kCGWindowBoundsinstead of assuming a fixed size. The post-resize capture also catches relayout bugs (AppKit reorders sublayers on the first layout pass).
Pixel probe — exact RGBA, no Python
osascript -l JavaScript -e '
ObjC.import("AppKit");
const rep = $.NSBitmapImageRep.imageRepWithContentsOfFile("/tmp/app-win.png");
function p(x,y){const c=rep.colorAtXY(x,y);return [c.redComponent,c.greenComponent,c.blueComponent,c.alphaComponent].map(v=>Math.round(v*255)).join(",");}
const h = rep.pixelsHigh*1, w = rep.pixelsWide*1;
JSON.stringify({w:w,h:h,center:p(Math.floor(w/2),Math.floor(h/2)),tl:p(2,2),bl:p(2,h-3),br:p(w-3,h-3)});'
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.
- 1mo ago First seen · 77 lines · 89 tokens per session scan A 2b3b23fe4d4d
gui-debug is a skill published in the GitHub repository natsukium/dotfiles (106 stars, last pushed today), licensed CC0-1.0. It adds 89 tokens to every session and 986 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
deep-research-team
This skill should be used when the user asks for "deep research", "research team", "comprehensive analysis", "research report", "investigate thoroughly", "compare X vs Y in depth", or needs synthesis across multiple sources with verification. It spawns a coordinated team of researcher agents across multiple rounds…
dustoff
Assess a project that's been dormant for months and produce a prioritized re-entry plan — lifecycle stage, what's stale or broken, and how far the project's tooling has drifted from current conventions — delegating every fix to the skill that owns it, and optionally capturing the plan as a tracking issue that survives…
qa-triage
Triage a QA-labeled report — investigate it against the code, classify it, and draft the technical issue(s) it warrants, stopping for approval before creating anything.
audit-permissions
This skill should be used when the user asks to "audit claude permissions", "audit permissions", "review local claude settings", "promote permissions to global", "clean up claude settings", "find permission patterns", or wants to identify project-local Claude Code permissions that should be added to global…
malo-find-skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
start
Enable auto TTS for this session.