Python 3.10.11 : python library-skills for your artificial intelligence.

The Python package library-skills is a lightweight command‑line toolkit designed to help developers build, test, and validate modular AI “skills.” These skills are small, self‑contained units of logic that can be executed independently or integrated into larger AI agents. The package focuses on simplicity, portability, and clear structure, making it useful for developers who want to experiment with tool‑calling systems or create custom capabilities for AI workflows.

A skill typically consists of a JSON descriptor and a Python function. The JSON file defines the skill’s name, description, and input schema, while the Python file contains the actual execution logic. This separation ensures that skills remain easy to document, validate, and reuse across different projects. With library-skills, developers can quickly inspect a skill’s schema, run it with custom input, or verify that its output matches the expected structure.

The command‑line interface provided by the package allows users to list installed skills, execute them directly, and validate input files without writing additional code. This makes the development cycle faster and more predictable. Instead of manually wiring functions together, developers can rely on a consistent interface that handles loading, parsing, and execution.

One of the main advantages of library-skills is its role in AI agent development. Modern agents often rely on tool‑calling, where the AI selects and triggers external functions based on user intent. Skills created with this package can be easily integrated into such agents, providing clear schemas and predictable behavior. This helps ensure that AI systems remain reliable, debuggable, and easy to extend.

Overall, library-skills is a practical utility for anyone building structured AI tools. It encourages clean design, modularity, and transparency, making it a valuable addition to Python environments focused on AI experimentation and agent development.

Let's install this python package.

python -m pip install library-skills
Collecting library-skills
  Downloading library_skills-0.0.19-py3-none-any.whl.metadata (5.4 kB)
...
Successfully installed library-skills-0.0.19 rich-toolkit-0.20.1 tomli-2.4.1

Let's make the first run:

library-skills.exe

 context
Project root               c:\lucru\PythonProjects
Target Python environment  not found

 Warning:  No target Python environment with site-packages or node_modules was found. Run from a
project root after installing dependencies, for example with 'uv sync' for Python or 'npm install' for
Node.js.

No installed or discovered skills found.
Copy the Library Skills tool skill into the project so agents know how to update, repair, and check
managed skills?
■ Copy Library Skills tool skill into the project so agents know how to update, repair, and check
managed skills? Copy Library Skills tool skill

 Target     Status               Path
 universal  tool skill: missing  .agents\skills\library-skills
 Copied:  library-skills (universal) -> .agents\skills\library-skills

Now you can create skills for your artificial intelligence:

Create a folder with two files:

my_skill/
    skill.json
    skill.py

First file named skill.json:

{
    "name": "hello_skill",
    "description": "Returnează un mesaj simplu",
    "input_schema": {
        "type": "object",
        "properties": {
            "name": {"type": "string"}
        },
        "required": ["name"]
    }
}

The second file named skill.py

def run(input):
    name = input["name"]
    return {"message": f"Salut, {name}!"}

Run the skill into your folder:

library-skills run my_skill --input '{"name": "Catalin"}'

See the schema:

library-skills schema my_skill

Validate the input:

library-skills validate my_skill input.json

List the skills:

library-skills list

This is all you need for a default basic skill with the library-skills.

Python 3.10.11 : CVSS (Common Vulnerability Scoring System) with cvss python package.

This Python package contains CVSS v2, v3 and v4 computation utilities and interactive calculator (for v2 and v3 only) compatible with Python 3. CVSS (Common Vulnerability Scoring System) is an standardized method for rating the severity of security issues on a scale from 0 (no impact) to 10 (critical).

See the pypi.org - cvss website.

Let's install the cvss python package.

python -m pip install cvss
Collecting cvss
  Downloading cvss-3.6-py2.py3-none-any.whl.metadata (3.8 kB)
Downloading cvss-3.6-py2.py3-none-any.whl (31 kB)
Installing collected packages: cvss
  WARNING: The script cvss_calculator.exe is installed in 'C:\python-3_10_11\Scripts' which is not on PATH.
  Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
Successfully installed cvss-3.6

How this works:

NVD Database (online)
        |
        |  JSON feed
        v
Python script ----> parses CVE + CVSS vector
        |
        |  uses cvss library
        v
Scores vulnerabilities (Base, Temporal, Environmental)
        |
        |  inserts results
        v
Your local database (SQL)
        |
        v
Dashboard / API / Alerts 

Simple code source example :

#!/usr/bin/env python3

# Demonstrates how to score a CVSS vector using the open-source "cvss" library.
# Validation and error handling included.

from cvss import CVSS3  # CVSS2, CVSS3, CVSS4 are available
import sys

def score_cvss_vector(vector: str):
    """
    Validates and scores a CVSS3 vector string.
    Returns scores and severities.
    """
    if not isinstance(vector, str) or not vector.strip():
        raise ValueError("Vector must be a non-empty string.")

    try:
        c = CVSS3(vector)
    except Exception as e:
        raise ValueError(f"Invalid CVSS3 vector: {e}")

    return c.clean_vector(), c.scores(), c.severities()

def main():
    if len(sys.argv) != 2:
        print("Usage: python cvss_score.py '<CVSS3_VECTOR>'")
        print("Example:")
        print("python main.py 'CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H'")
        sys.exit(1)

    vector = sys.argv[1]

    try:
        clean_v, scores, severity = score_cvss_vector(vector)
        print("Input vector:", vector)
        print("Normalized vector:", clean_v)
        print("Scores:", scores)
        print("Severity:", severity)
    except ValueError as e:
        print("Error:", e)
        sys.exit(1)

if __name__ == "__main__":
    main()

The result is this:

python main.py CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Input vector: CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Normalized vector: CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Scores: (9.8, 9.8, 9.8)
Severity: ('Critical', 'Critical', 'Critical')

Another source code with examples:

#!/usr/bin/env python3
# Requires: pip install cvss

from cvss import CVSS3

# Example vulnerabilities (safe, educational)
vulns = [
    {
        "language": "Python",
        "title": "Unsafe eval usage",
        "description": "Code that executes user-provided input using eval().",
        "cvss_vector": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N"
    },
    {
        "language": "C#",
        "title": "Insecure deserialization",
        "description": "BinaryFormatter deserialization of untrusted data.",
        "cvss_vector": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H"
    },
    {
        "language": "Godot Engine",
        "title": "Unvalidated file path access",
        "description": "Loading files from paths provided by the user without validation.",
        "cvss_vector": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N"
    }
]

def analyze_vulnerabilities(vuln_list):
    for v in vuln_list:
        print("\n====================================")
        print("Language:", v["language"])
        print("Issue:", v["title"])
        print("Description:", v["description"])
        print("CVSS Vector:", v["cvss_vector"])

        try:
            cv = CVSS3(v["cvss_vector"])
            base, temp, env = cv.scores()
            sev_base, sev_temp, sev_env = cv.severities()

            print("Base Score:", base, "-", sev_base)
            print("Temporal Score:", temp, "-", sev_temp)
            print("Environmental Score:", env, "-", sev_env)

        except Exception as e:
            print("Invalid CVSS vector:", e)

if __name__ == "__main__":
    analyze_vulnerabilities(vulns)

This is the result:

python main_002.py

====================================
Language: Python
Issue: Unsafe eval usage
Description: Code that executes user-provided input using eval().
CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N
Base Score: 8.1 - High
Temporal Score: 8.1 - High
Environmental Score: 8.1 - High

====================================
Language: C#
Issue: Insecure deserialization
Description: BinaryFormatter deserialization of untrusted data.
CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Base Score: 9.8 - Critical
Temporal Score: 9.8 - Critical
Environmental Score: 9.8 - Critical

====================================
Language: Godot Engine
Issue: Unvalidated file path access
Description: Loading files from paths provided by the user without validation.
CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N
Base Score: 4.4 - Medium
Temporal Score: 4.4 - Medium
Environmental Score: 4.4 - Medium

Python 3.10.11 : Show the CVE's results with opencve token.

Today, this simple source code use token from opencve.io - website to show CVE's results.

import requests

API_URL = "https://app.opencve.io/api/cve"
TOKEN = "opc_org.<token_id>.<secret>"

headers = {
    "Authorization": f"Bearer {TOKEN}",
    "Accept": "application/json",
}

params = {
    "vendor": "microsoft",
    "cvss": "critical",
    "page": 1,
}

resp = requests.get(API_URL, headers=headers, params=params)
data = resp.json()

for cve in data["results"]:
    print(cve["cve_id"], cve["description"])

This is the result:

python main_001.py
CVE-2026-58289 Access of resource using incompatible type ('type confusion') in Microsoft Edge (Chromium-based) allows an unauthorized attacker to execute code over a network.
CVE-2026-45499 Server-side request forgery (ssrf) in Azure OpenAI allows an authorized attacker to elevate privileges over a network.
CVE-2026-41106 Url redirection to untrusted site ('open redirect') in M365 Copilot allows an unauthorized attacker to elevate privileges over a network.
CVE-2026-57100 Server-side request forgery (ssrf) in Microsoft Entra Provisioning Service (SyncFabric) allows an authorized attacker to elevate privileges over a network.
CVE-2026-54130 Missing authentication for critical function in M365 Copilot allows an unauthorized attacker to disclose information over a network.
CVE-2026-48584 Execution with unnecessary privileges in Azure Synapse allows an authorized attacker to elevate privileges over a network.
CVE-2026-45480 Improper authentication in Azure Active Directory allows an unauthorized attacker to elevate privileges over a network.
CVE-2025-62821 Microsoft HEIF Image Extensions 1.2.22.0 has an out-of-bounds read because CHEIFItemInfoEntry_GetDataSize can return success while leaving the reported data size as 0. This causes a caller to make a 1-byte allocation. Later, CopyPixels computes copy_size = stride * abs(roi_height) but does not check the source buffer length before a memmove call.
CVE-2026-47647 Improper access control in Microsoft Dynamics 365 allows an authorized attacker to elevate privileges over a network.
CVE-2026-48582 Missing authorization in Microsoft Exchange Online allows an authorized attacker to elevate privileges over a network.