📦

vision

Name: vision
Author: kotot

Révision du contenu r1 Risque élevé ⚙️ Commandes externes🌐 Accès réseau📁 Accès au système de fichiers🔑 Variables d’environnement

Analyze Images with a Vision Model

Text-only agents cannot inspect screenshots, diagrams, or scanned pages. This skill sends one image and a focused prompt to your configured vision model.

Prend en charge: Claude Codex Code(CC)

⚠️ 38 Médiocre

Maintain this Skill? Claim it ⚠️ 38 Médiocre

Arrêtez-vous et demandez confirmation avant l’installation.

Examinez le plan et obtenez le consentement explicite de l’utilisateur avant de modifier des fichiers.

Installer avec mon Agent

Copiez cette demande dans votre Agent. Elle contient la page Skill de référence et le manifeste.

Demande d’agent

Review the Skillstore skill "vision" from https://skillstore.io/skills/kotot-vision.md and its manifest at https://skillstore.io/api/skills/kotot-vision/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Votre Agent doit toujours présenter son plan et demander toute confirmation exigée par la politique de sécurité.

Ressources lisibles par les agents

Utilisez ces liens lorsqu’un Agent IA, un crawler ou un script a besoin d’un contexte propre au lieu de lire toute la page.

Détail Markdown GET /skills/kotot-vision.md manifest signé GET /api/skills/kotot-vision/manifest lockfile signé GET /api/skills/kotot-vision/lockfile

Tester

Utilisation de « vision ». Describe a dashboard screenshot with a failed deployment.

Résultat attendu:

The deployment status is Failed. The visible error cites a missing database connection variable in the production environment.

Utilisation de « vision ». Transcribe a photographed receipt.

Résultat attendu:

Merchant: North Market
Date: 2026-08-14
Total: $42.80
The item description on the second line is unclear.

Utilisation de « vision ». Explain an architecture diagram.

Résultat attendu:

The client sends requests through an API gateway to two services. Both services share a database, while the worker also reads from a queue.

Audit de sécurité

Risque élevé

v1 • 16/08/2026 Ouvrir le rapport versionné

The two outbound HTTP findings are confirmed because the script posts image data and prompts to a configurable endpoint. Most other findings are documentation, placeholders, or restricted configuration access, but inherited project settings can silently redirect uploads.

Fichiers analysés

382

Lignes analysées

Éléments à examiner

Faux positifs ignorés

Préoccupations de sécurité confirmées (1)

Élevé

scripts/see.py:56-70 scripts/see.py:172-182 scripts/see.py:208-242 SKILL.md:69-73

Project Configuration Can Redirect Image Uploads

The script searches parent directories for project settings and trusts their VISION_BASE_URL. A repository can redirect selected images and prompts to its endpoint.

The code directly combines upward project configuration discovery with automatic POST transmission to the resolved endpoint.

Éléments d’évaluation des capacités (2)

Il s’agit de capacités locales réelles qui peuvent être attendues pour cette compétence ; elles nécessitent donc un examen, mais ne sont pas comptabilisées comme comportement malveillant confirmé.

Faible

scripts/see.py:240

Python HTTP libraries

req = urllib.request.Request(base_url, data=data, headers=headers, method="POST")

The script constructs a POST request to a configurable endpoint with the image payload, prompt, and optional authorization header.

Faible

scripts/see.py:242

Python HTTP libraries

with urllib.request.urlopen(req, timeout=timeout) as resp:

urlopen executes the outbound request, transmitting image content and prompts outside the local process.

Facteurs de risque

⚙️ Commandes externes (27)

🌐 Accès réseau (11)

scripts/see.py:37 scripts/see.py:240 scripts/see.py:242 scripts/see.py:12 scripts/see.py:25 scripts/see.py:197 scripts/see.py:206 SKILL.md:59 SKILL.md:77 SKILL.md:91 SKILL.md:101

📁 Accès au système de fichiers (15)

scripts/see.py:200 scripts/see.py:8 scripts/see.py:51 scripts/see.py:102 scripts/see.py:103 scripts/see.py:8 scripts/see.py:51 scripts/see.py:100 scripts/see.py:102 scripts/see.py:103 SKILL.md:86 SKILL.md:73 SKILL.md:96 SKILL.md:73 SKILL.md:96

🔑 Variables d’environnement (14)

scripts/see.py:53 scripts/see.py:164 scripts/see.py:178 scripts/see.py:100 scripts/see.py:107 scripts/see.py:177 scripts/see.py:15 scripts/see.py:187 scripts/see.py:200 scripts/see.py:237 scripts/see.py:238 SKILL.md:79 SKILL.md:93 SKILL.md:103

Audité par: codex

Partager et citer ce rapport

Partagez le rapport d’évaluation versionné, le badge neutre, la carte intégrable et les citations. Skillstore présente les preuves sans décider si ce Skill est sûr.

Ouvrir le rapport versionné

Copier le lien du rapport

https://skillstore.io/skills/kotot-vision/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Badge Markdown

[![Skillstore security assessment](https://skillstore.io/badges/skills/kotot-vision/security.svg)](https://skillstore.io/skills/kotot-vision?utm_source=security_passport_badge)

Badge HTML

<a href="https://skillstore.io/skills/kotot-vision?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/kotot-vision/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Carte intégrable

<iframe src="https://skillstore.io/embed/skills/kotot-vision.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Citations académiques (APA · BibTeX · CFF)

Citation APA

kotot. (2026). vision security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/kotot-vision/audits/1

Citation BibTeX

@techreport{kotot-kotot-vision-2026,
  author = {kotot},
  title = {vision security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/kotot-vision/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "vision security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "kotot"
date-released: "2026-08-16"
url: "https://skillstore.io/skills/kotot-vision/audits/1"
identifiers:
  - type: other
    value: "skillstore:kotot-vision:audit:1"
    description: "Skillstore immutable audit report identifier"

Comparer les variantes

3 variantes installables

Chaque auteur reste un skill installable distinct. La variante recommandée est classée selon les preuves Skillstore.

Pourquoi cette variante est première

Meilleur score Skillstore

Auteur / source Score Skillstore Niveau de confiance des preuves Utilisation Skillstore Mis à jour

0xSero Recommandé

0xsero-vision

Score Skillstore 78

Niveau de confiance des preuves Élevé

Utilisation Skillstore 128

Mis à jour

2026-10-12

verging-ai

verging-ai-vision

Score Skillstore 50

Niveau de confiance des preuves Élevé

Utilisation Skillstore 50

Mis à jour

2026-09-09

kotot Actuelle

kotot-vision

Score Skillstore 38

Niveau de confiance des preuves Moyen

Utilisation Skillstore 40

Mis à jour

2026-09-09

Score Skillstore

Pourquoi ce score Niveau de confiance des preuves: Moyen

Architecture

Maintenabilité

Contenu

Communauté

Conformité aux spécifications

Ce que vous pouvez créer

Diagnose Screenshot Errors

Read visible error messages and identify likely causes from application screenshots.

Extract Scanned Text

Transcribe text from receipts, forms, and scanned pages while preserving useful structure.

Interpret Visual Research

Summarize charts and diagrams, including labels, values, relationships, and visible caveats.

Essayez ces prompts

Describe an Image

Analyze [image path or URL]. Describe the main subjects, setting, visible text, and important details.

Transcribe Visible Text

Read [image path or URL]. Transcribe all visible text and preserve headings, lists, and table relationships.

Diagnose a Screenshot

Inspect [screenshot path]. Identify the exact error, affected component, likely cause, and evidence visible in the interface.

Extract and Compare Evidence

Analyze [image path or URL] against [expected state]. List matches, discrepancies, uncertain details, and evidence for each conclusion.

Bonnes pratiques

Use a specific question that names the details you need.
Verify the resolved endpoint before sending confidential images.
Ask the model to mark unreadable or uncertain content.

Éviter

Do not send secrets or regulated data to an untrusted endpoint.
Do not treat OCR output as authoritative without verification.
Do not request broad descriptions when exact values are required.

Foire aux questions

Which image sources are supported?

The script accepts one local file path or one HTTP or HTTPS image URL.

Does this skill require a specific provider?

No. It requires an OpenAI-compatible chat completions endpoint with vision support.

Is an API key required?

Remote services usually require one. Local endpoints may accept an empty key.

Where is configuration loaded from?

Environment variables have priority. The script can also read Claude settings and OpenClaw dotenv files.

Can I inspect a request before sending it?

Yes. Use dry-run mode to preview the destination and request structure without transmitting the image.

Does the skill guarantee accurate OCR?

No. Accuracy depends on image quality and the selected model. Verify important text and numerical values.

Détails du développeur

Auteur

kotot

Licence

MIT

Révision Skillstore

Avis de version

L’auteur n’a déclaré aucune version.

Dépôt

https://github.com/kotot/vision/tree/1740cbd119cd11cfacf51d29fe911d8c4ba4cf85/vision

Réf

d7ec88d380bc0952a475899451d682383ab5d76c

Récence de maintenance

16/08/2026

Utilisation

0 téléchargements · 0 vues

Structure de fichiers

📁 scripts/

📄 see.py

📄 SKILL.md