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 calesthio/generative-media-skills --skill google-cloud-visiongit clone --depth 1 https://github.com/calesthio/generative-media-skillsWrote 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/calesthio/generative-media-skills/google-cloud-vision)<a href="https://agentmods.dev/skills/calesthio/generative-media-skills/google-cloud-vision"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/google-cloud-vision/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/calesthio/generative-media-skills/google-cloud-vision"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/google-cloud-vision.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 | $0.00096 | $0.06026 |
| Opus 5 | $0.00048 | $0.03013 |
| Sonnet 5 | $0.00019 | $0.01205 |
| Haiku 4.5 | $0.00010 | $0.00603 |
Grade A, and why
google-cloud-vision 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.
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 — 335 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Google Cloud Vision API for still-image understanding
Use Google Cloud Vision when the production task needs structured annotations from still images: general labels, object boxes, text/OCR, dense document text, explicit-content likelihoods, dominant colors, crop suggestions, or web-reference signals. Treat it as a computer-vision annotation service, not as a conversational image reasoner.
Verified on 2026-07-11 against first-party Google Cloud documentation whose Cloud Vision pages were last updated 2026-07-07 unless otherwise noted. Pricing, quotas, endpoints, product lifecycle, supported features, and model behavior are volatile; recheck the linked Google pages before promising a production SLA, budget, data-region commitment, or migration path.
Use and non-use boundaries
Documented fact: Cloud Vision API feature types include TEXT_DETECTION, DOCUMENT_TEXT_DETECTION, LABEL_DETECTION, OBJECT_LOCALIZATION, SAFE_SEARCH_DETECTION, IMAGE_PROPERTIES, CROP_HINTS, and WEB_DETECTION, among other features such as landmarks, logos, and faces. Source: Google Cloud Vision features list, verified 2026-07-11.
Use Cloud Vision for:
- Still images that need stable JSON fields rather than prose: labels with scores, object bounding polygons, OCR structure, SafeSearch likelihoods, dominant color values, crop vertices, web matches, or batch outputs.
- Backend media pipelines where images live in Cloud Storage and annotations must be stored, audited, thresholded, or joined with internal records.
- High-throughput OCR triage, moderation prefilters, asset tagging, crop automation, or web-reference discovery where deterministic fields and quotas matter more than open-ended reasoning.
Do not use Cloud Vision as the primary tool for:
- Gemini multimodal reasoning: use Gemini/Vertex AI image understanding when the user asks for reasoning, visual question answering, comparison across many images, narrative descriptions, explanation, or flexible prompt-following. Documented distinction: Gemini image understanding accepts image-plus-text prompts and has limitations around precise spatial localization and possible hallucination; Cloud Vision returns feature-specific annotations. Source: Google Cloud Gemini image understanding docs, last updated 2026-07-09, verified 2026-07-11.
- Video analysis: use Video Intelligence API or another video-understanding provider for per-shot, per-frame, segment-level, object tracking, text in video, explicit content in videos, audio transcription, or live-stream analysis. Source: Google Cloud Video Intelligence overview, verified 2026-07-11.
- Image generation or editing: Cloud Vision analyzes images; it does not generate, retouch, upscale, remove backgrounds, or alter pixels.
- Custom category training: use AutoML Vision or current Vertex AI/AutoML routes when the business needs a custom labeler/object detector. Cloud Vision's standard features are fixed provider models. The Product Search docs also state that AutoML Vision enables custom image labeling; verify current AutoML/Vertex recommendations before starting a new custom-model project.
- Human reference analysis, provenance, or legal authenticity:
WEB_DETECTIONcan return matching pages/images and inferred web entities, but it does not prove authorship, license, consent, manipulation history, source-of-truth identity, or rights clearance. Use it as evidence collection for a human reviewer, not as the reviewer.
What ships with it
1 file 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.
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.
- 9d ago First seen · 335 lines · 0 tokens per session scan A 6e9b4e464156
google-cloud-vision is a skill published in the GitHub repository calesthio/generative-media-skills (171 stars, last pushed 2mo ago), licensed MIT. It adds 96 tokens to every session and 6,026 once invoked, about $0.0005 per session on Opus 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 2026-09-03.
Other skills, from other repositories
cliptalk-cover-director
Produces evidence-backed cover candidates and reviewable cover variants for a ClipTalk video. Use when the user asks for a cover, poster frame, thumbnail, or multiple cover directions; do not use for timeline editing or social-video reframing.
cliptalk-smart-reframe
Creates a subject-aware, time-varying crop track and a review-only social-format preview from an accepted ClipTalk cut. Use for automatic vertical, square, or portrait reframing; do not use for a fixed manual crop or before content editing is accepted.
cliptalk-content-extractor
Locates and assembles source passages matching a semantic request. Use for extracting explanations, topics, quotes, demonstrations, or other specifically described content.
cliptalk-interview-editor
Produces a coherent interview edit by combining speaker discovery, topic selection, dialogue context, cleanup, subtitles, and preview. Use for interviews, podcasts, testimonials, or question-and-answer recordings.
cliptalk-shortform-hook-director
Finds and assembles a reviewable short-form cut with a strong opening hook. Use for Shorts, Reels, social clips, talking-head cutdowns, or requests for a punchier opening.
cliptalk-social-reframe-exporter
Creates a review-only 9:16, 4:5, 1:1, or 16:9 version from an existing accepted ClipTalk cut, then checks the rendered preview. Use only when a cut already exists and the user asks to adapt it for Shorts, Reels, Douyin, Xiaohongshu, WeChat Channels, or square feeds; do not use when the user still needs content found…