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 teimurjan/blazediff --skill image-comparegit clone --depth 1 https://github.com/teimurjan/blazediffWrote 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/teimurjan/blazediff/image-compare)<a href="https://agentmods.dev/skills/teimurjan/blazediff/image-compare"><img src="https://agentmods.dev/badge/skills/teimurjan/blazediff/image-compare/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/teimurjan/blazediff/image-compare"><img src="https://agentmods.dev/badge/skills/teimurjan/blazediff/image-compare.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 | $0.00049 | $0.02376 |
| Opus 5.5 | $0.00020 | $0.00950 |
| Sonnet 5.5 | $0.00010 | $0.00475 |
| Haiku 4.5 | $0.00005 | $0.00238 |
Grade A, and why
image-compare 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 — 184 lines — stays where its author put it; the contents beside it link to each section on GitHub.
image-compare
Two image files in, a verdict out. No baselines, no manifest, no browser.
CLI binary is blazediff-cli (npm i -g @blazediff/cli).
Not this skill: capturing a running app, managing baselines, or re-running a
visual regression suite → use the blazediff skill instead (it owns
.blazediff/, blazediff-agent, and the check/rewrite loop). If the working
directory has a .blazediff/manifest.json and the user is asking about routes
rather than files, you are in the wrong skill.
Be terse
- Pass
--jsonon every call; parse fields. Do not echo CLI output. - Never send whole images to a model when
interprethas located the regions — see Send crops, not pages below. This is the point of the skill. - One final summary line:
N regions | <severity> | <one-clause verdict>.
Pick the command by the question
| The user is asking | Command | Read from it |
|---|---|---|
| "what changed?" / "describe the differences" | interpret |
regions[] — bbox, changeType, position |
| "are these identical?" | core-native |
different pixels (0 = byte-identical after decode) |
| "how different, perceptually?" | ssim |
SSIM score, 1.0 = identical |
| "did quality degrade?" (compression, resize, generative output) | msssim |
multi-scale score, 1.0 = identical |
| "where are the structural differences?" | gmsd -o map.png |
deviation, 0.0 = identical |
interpret is the default. The others answer how much; only interpret
answers what.
blazediff-cli interpret a.png b.png --json
Exit codes: 0 no regions, 1 regions found, 2 error. Safe to branch on.
Picking --source
pixel(default) — pixel-aligned renders: screenshots, PNG exports, chart renders, PDF rasterizations. Anything produced twice by the same renderer.ms-ssim/ssim— images with resampling or compression noise: JPEGs, camera captures, generative-model output, anything that went through a lossy round-trip. Per-pixel deltas are everywhere in these, sopixelreturns one giant region; the metric sources threshold a similarity map instead. Their boxes are coarser (map-grid resolution), but the pixel counts inside them are still exact.
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 · 184 lines · 49 tokens per session scan A 63a604644508
image-compare is a skill published in the GitHub repository teimurjan/blazediff (352 stars, last pushed today), licensed MIT. It adds 49 tokens to every session and 2,376 once invoked, about $0.0002 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
a11y-ally
Use when running comprehensive WCAG accessibility audits with axe-core + pa11y + Lighthouse, generating context-aware remediation, or testing video accessibility. Supports 3-tier browser cascade with graceful degradation.
accessibility-testing
WCAG 2.2 compliance testing, screen reader validation, and inclusive design verification. Use when ensuring legal compliance (ADA, Section 508), testing for disabilities, or building accessible applications for 1 billion disabled users globally.
frontend-manual-testing
Use when the task changes anything a user sees in a web app - booting it, driving flows with the browser tools, the four-width check, console/network/keyboard checks and the visual checklist.
playwright-scanner
Internal helper: behavioral scans via Playwright for keyboard and state.
visual-proof
Capture before/after screenshot evidence of a web page at real breakpoints (375, 768, 1440) in both light and dark themes, using Playwright against a local dev server or a deployed URL. Use before starting any visual change to capture the current state, and again after to prove the delta. Also use when asked to "show…
visual-qa-testing
Visually QA a web application by launching it in Cursor's built-in browser, taking screenshots, checking console errors, and auditing network requests. Use after making UI changes to verify they look correct.