← Plugin catalog
Security

VisionEdge AIScan

VisionEdge, s. r. o. v1.7.8

Security checks for AI code

Create a free account to save this plugin.

From the developer

Declared capabilities

  • Detect hallucinated, malicious and typosquatted npm and PyPI packages before installation
  • Find known vulnerabilities in dependencies with OSV, GitHub advisories, CISA KEV and EPSS
  • Find hardcoded and browser exposed API keys, always reported redacted

Declared limitations

  • Must not attack or break into systems. The plugin only runs passive checks on sites the user confirms they own, and offers no exploitation.
  • Must not collect or keep credentials. The plugin never asks for or stores API keys.
  • Must not create malware. The plugin detects malicious packages and does not build them.

Not specified in the available declarations: pricing & account requirements.

Onboarding example

Is the npm package react-codeshift safe to install?

View onboarding source →

Sources & dates

Developer declarations; not independently tested. Listing last checked Nov 16, 2026. Package observed Nov 16, 2026.

View package source →

Description

From the marketplace listing

AIScan checks software written with AI coding tools for the security mistakes those tools often make. What you can ask: - Is a package safe to install? AIScan catches packages that do not exist (AI hallucinations that attackers register with malware, known as slopsquatting), known malicious packages, brand new look-alike packages, typosquats and known vulnerabilities from OSV and GitHub advisories, with the CISA known exploited flag and EPSS exploit probability. - Check a whole package.json, requirements.txt, pyproject.toml or lockfile and get the exact upgrades. - Scan code you share or that your AI assistant wrote: API keys in code or exposed to the browser (always reported redacted), model output reaching eval, a shell, SQL or HTML, user input inside system prompts, public LLM endpoints without login or rate limits, Supabase tables without Row Level Security, open Firebase rules, risky MCP servers and agent permissions, and hidden Unicode instructions in rules files. - Scan a public GitHub repository end to end. - Run a passive check of a website you own: HTTPS, security headers, keys inside the JavaScript sent to visitors, public source maps, Supabase and Firebase setup. - See the newest vulnerabilities in AI frameworks, LLM SDKs, vector databases and MCP tooling. Every finding has a severity, a plain language explanation, the location, a concrete fix and its mapping to the OWASP Top 10 for LLM Applications, the OWASP Top 10 for Agentic Applications and CWE. Code is processed only to answer the request and is not stored. Limits: AIScan is an automated scanner and cannot prove that software is secure. Website checks are passive and need the user to confirm they own the site or may test it. Deep website checks and live monitoring are available on the AIScan website.

Language: English · Automatically detected from descriptions.

Changes

No changes recorded yet. Follow this plugin to keep track of future changes.

First seen Nov 16, 2026. Last seen Nov 16, 2026.

Keywords

Files & skills

File archives

Plugin package11 files · 5.37 KBBrowse files →
Skill instructions
aiscan-onboarding899 Bytes

View saved version →

---
name: aiscan-onboarding
description: Introduce VisionEdge AIScan to a new user in a few sentences and suggest a first check. Use it the first time the user opens the plugin or asks what AIScan can do.
---

# Getting started with AIScan

Explain in two or three short sentences what AIScan does: it checks software written with AI for hallucinated or malicious packages, known vulnerabilities, leaked keys, unsafe use of model output, open databases and risky agent settings.

Then offer three ways to start and let the user pick one:

1. "Paste your package.json or requirements.txt and I will check every dependency."
2. "Give me a public GitHub repository link and I will scan the whole project."
3. "Name a package you are about to install and I will tell you whether it is safe."

Keep it short. Do not list every rule. Do not ask for API keys, passwords or tokens: AIScan never needs them.
secure-ai-coding2.67 KB

View saved version →

---
name: secure-ai-coding
description: Check packages, code, repositories and websites for the security mistakes typical of software written with AI, using the VisionEdge AIScan tools. Use it before installing any dependency, after writing code that handles LLM output, secrets, databases or agent tools, and before deploying.
---

# Secure AI coding with AIScan

AIScan finds problems that AI coding tools often introduce: invented package names, vulnerable versions, leaked keys, unsafe use of model output, open databases and risky agent settings.

## When to call which tool

- **Before adding or installing any package** (npm install, pip install, editing package.json, requirements.txt or pyproject.toml): call `check_package` for every new package.
  - `do_not_install`: do not install it. Tell the user why and look for the real package name.
  - `caution`: explain the signal (very new, little used, look-alike name) and confirm with the user.
  - `vulnerable_version`: use the fixed version the tool returns.
- **A dependency file is shared or the user asks whether dependencies are safe**: call `check_dependencies` with the file name and full content.
- **After writing or changing code** that calls an LLM, renders model output, builds SQL or shell commands, handles API keys, defines Supabase or Firebase rules, or configures MCP servers and agents: call `scan_code` with the changed files and their real relative paths. Paths matter (for example `app/api/chat/route.ts` is a server route, `components/Chat.tsx` runs in the browser).
- **A public GitHub repository**: call `scan_github_repo`.
- **A website**: first ask the user to confirm they own it or may test it. Only then call `scan_website` with `confirm_authorized` set to true.
- **Questions about new vulnerabilities in AI frameworks, SDKs or MCP tooling**: call `get_ai_threat_feed`, optionally with a package name.

## How to present results

1. Start with the grade and the number of critical and high findings.
2. For each critical and high finding give one sentence on what is wrong, where it is, and the exact fix.
3. Offer to apply the fixes. When you change code, scan it again to confirm the fix.
4. Never print secret values. Refer to a leaked key by its location and tell the user to rotate it at the provider.
5. Mention the remaining medium and low findings briefly.

## Limits

- Website checks are passive. Deep checks (exposed files, database read access, CORS) and live monitoring of new vulnerabilities are available on the AIScan website after domain verification.
- This is an automated scan. It finds common problems but cannot prove that software is secure.
- Never use these tools to test systems the user does not own or may not test.

Release notes

First release: package safety checks, dependency file checks, code scanning, public GitHub repository scanning, passive website checks and the AI threat feed.

Declared in the saved package. Remote tools may change independently.

Package details

Publisher declarations from the archived package. These are separate from our research and the live service's terms.

Package license
Proprietary
Package author
VisionEdge, s. r. o.
Keywords
See keywords
Declared availability
No country restrictions declaredDeclared regions; live availability may differ.
Getting started skill
./skills/aiscan-onboarding/SKILL.md
Commerce declaration
Does not support commerceThis does not establish whether access is free or paid.
Review scenarios
5 positive · 3 negativeScenario counts, not test results.

Declared capabilities

  • Detect hallucinated, malicious and typosquatted npm and PyPI packages before installation
  • Find known vulnerabilities in dependencies with OSV, GitHub advisories, CISA KEV and EPSS
  • Find hardcoded and browser exposed API keys, always reported redacted
  • Find unsafe handling of LLM output: eval, shell, SQL, HTML rendering
  • Check Supabase Row Level Security and Firebase rules
  • Review MCP server and AI agent configurations for excessive permissions
  • Detect hidden Unicode instructions in agent rules files
  • Passive security check of a website the user owns
  • Newest vulnerabilities in AI frameworks and MCP tooling

Some manifest fields differ or could not be read. The structured report retains the source references.

Package observed Nov 16, 2026.

Technical details
First seen
Nov 16, 2026 · 18:00 UTC
Last seen
Nov 16, 2026 · 18:00 UTC
Collection status
Collected

plugin_asdk_app_6af613dd5da881919a31ce59c0dd6816

Download plugin data (JSON)

Plugin details: served from cache. Prepared . Refreshed when plugin data changes or after 6 hours. Your account and Follow status are live.