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 buildinternet/uploads --skill annotate-screenshotsgit clone --depth 1 https://github.com/buildinternet/uploadsWrote 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/buildinternet/uploads/annotate-screenshots)<a href="https://agentmods.dev/skills/buildinternet/uploads/annotate-screenshots"><img src="https://agentmods.dev/badge/skills/buildinternet/uploads/annotate-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.
<a href="https://agentmods.dev/skills/buildinternet/uploads/annotate-screenshots"><img src="https://agentmods.dev/badge/skills/buildinternet/uploads/annotate-screenshots.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Prompt Injection · line 55 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Prompt Injection · line 57 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
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 | $0.00169 | $0.01800 |
| Opus 5.5 | $0.00068 | $0.00720 |
| Sonnet 5.5 | $0.00034 | $0.00360 |
| Haiku 4.5 | $0.00017 | $0.00180 |
Grade A, and why
annotate-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 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 — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Annotating screenshots
When to annotate
Add annotations when a screenshot needs to draw the reader's eye somewhere specific — a new element, a bug, a diff — rather than leaving them to find it in the caption. A plain before/after pair is often enough; reach for annotation when "look at the top-right corner" is easier to say with an arrow than with words, or when a screenshot contains a secret that must be hidden before it's shared anywhere (uploads.sh URLs are public).
Two commands, same underlying spec format:
| Command | Input | Targeting |
|---|---|---|
uploads screenshot <url|file.html> --annotate <spec> |
captures a live page | CSS selectors (preferred) or pixel coordinates |
uploads annotate <image> --spec <spec> |
an existing PNG/JPEG file | pixel coordinates only — selectors are rejected |
Prefer selectors, resolved live, over guessing pixel coordinates. An
agent estimating x/y/w/h by eye is unreliable — it's easy to point at
the wrong element or clip a box awkwardly. uploads screenshot --annotate
resolves a CSS selector against the real DOM at capture time (local backend
only in v1 — --via remote rejects selector-bearing specs up front), so use
it whenever you're capturing the page yourself. Fall back to pixel
coordinates with uploads annotate only when you're marking up an image you
already have and there's no live page to query — a screenshot someone else
took, a photo, a diagram exported from elsewhere.
The spec format
A spec is a JSON document: { "version": 1, "annotations": [...] }.
annotations is a non-empty array; each entry has a type plus fields for
that type. House style (color, stroke weight, sketchy roughness, font) is
fixed and not configurable — the one style knob is a per-annotation color
override.
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 · 161 lines · 169 tokens per session scan A e72a922bb910
annotate-screenshots is a skill published in the GitHub repository buildinternet/uploads (19 stars, last pushed today), licensed Apache-2.0. It adds 169 tokens to every session and 1,800 once invoked, about $0.0007 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
design-qa
Deterministic design review for UI/UX on web and terminal screens: measure contrast, spacing, overlap, clipping and copy defects, then judge the rest. Use when reviewing a screen, dashboard, or interface's visual design.
aiox-ux-design-expert
Activate Uma (ux-design-expert) for UX/UI Designer & Design System Architect. Complete design workflow - user research, wireframes, design systems, token extraction, component building, and quality assurance.
aiox-ux-design-expert
UX/UI Designer & Design System Architect (Uma). Complete design workflow - user research, wireframes, design systems, token extraction, component building, and quality assurance.
hatch3r-design-system-detect
Detects existing design tokens, component library, and theming convention in a project before authoring new UI primitives — output a concise inventory for downstream implementers.
accessibility-compliance
Implement WCAG 2.2 compliant interfaces with mobile accessibility, inclusive design patterns, and assistive technology support. Use when auditing accessibility, implementing ARIA patterns, building for screen readers, or ensuring inclusive user experiences.
frontend-taste
Premium frontend design skills that override LLM defaults. Prevents generic 'AI-looking' UI. Enforces deterministic typography, calibrated color, asymmetric layouts, spring physics, hardware-accelerated motion. Triggers: 'UI tasarla', 'design a UI', 'premium landing page', 'frontend taste', 'anti-slop', 'dashboard…