OpenViSTAI · Agents · Blockchain · Crypto
Menu
Agent Frameworks

Agent Skills Platform

Agent Skills Platform is a Python developer tool for creating tested, installable AI-agent skills from real workflows and supporting evidence, then governing their publication and lifecycle through a team marketplace.

Independent OpenViST project profile, based on linked sources.

Agent Skills Platform — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when recurring work relies on expert decisions and evidence that should be captured, tested and reused by an AI agent. Choose the private path for individual use; use a governed marketplace when teammates will install or reuse the skill itself.

Who it is for

  • Subject-matter experts who want to capture recurring work as reusable agent skills.
  • Developers or creators who build and test skill packages.
  • Marketplace operators who manage team approval, release, distribution, updates, quarantine and rollback.
  • Teammates who install approved skills through a desktop app or supported agent tools.

Documented uses

  • Turn examples and a prompt into a reusable workflow skill.
  • Package instructions, scripts when needed, evaluations and security checks for review.
  • Publish and distribute approved skills through a governed team marketplace.
  • Maintain skills using classified run evidence and separately validated changes.
  • Browse and install team skills through the described desktop app.
Before you choose

Requirements & limits

The repository is public and identifies an MIT license. The README describes a desktop app for teammate installation and a separate operator process for marketplace administration.

  • A skill can include instructions, functional scripts when needed, evals, security checks and a representative-run record.
  • The team marketplace path is intended for cases where teammates will install or reuse the skill; private creation does not require marketplace setup.
  • Marketplace setup uses a governed GitHub or GitLab library, according to the README.
  • The README describes distribution across 17 platforms and names Claude Code, Codex, Cursor and GitHub Copilot among relevant ecosystems.
  • Run evidence may inform draft patterns, but a separately validated change is required to update the executable skill; the README explicitly distinguishes this from autonomous self-modification.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Create and privately verify an agent skill from an existing workflow

Turn a recurring workflow and its examples into a reusable skill, then verify it and install it privately.

Before you start

  • Access to an AI agent where the platform creator is installed.
  • A workflow description and examples of the work, such as reports or spreadsheets.
  • For workflows using an API, database, MCP, codebase, or structured file, Semantic Recon runs automatically before implementation; a deliberately local, source-free installation can use the documented opt-out.

Steps

  1. Open the AI agent you already use and attach examples of the workflow.

  2. Paste a request beginning with /agent-skills-platform, describe the workflow, and state its decision criteria and constraints.

  3. Answer the creator’s questions about business decisions, then review the representative result and correct it until it is right.

  4. For a private skill, create, verify, and install it privately; no marketplace setup is needed.

Expected outcome

A reusable, installable skill package with instructions and, when needed, functional scripts, evals, security checks, and a representative-run record.

Guide limitations

  • The creator must be installed in the agent; otherwise, the instructions say to send the workflow section to a marketplace operator.
  • The skill’s business decisions require authorization from the workflow owner.
  • The documented Semantic Recon opt-out applies only to a deliberately local, source-free installation.
Source-based practical guide

Install a team skill using the desktop app

Let a teammate install a skill from a team-library link using the desktop app.

Before you start

  • The team member has access to a team-library link.
  • The desktop app is available to the team member.

Steps

  1. Open the desktop app.

  2. Paste the team-library link.

  3. Click Install.

  4. Choose GitHub Copilot or another tool offered by the app.

Expected outcome

The selected skill is installed in the chosen tool, with no terminal, Python, or Git required for the teammate.

Guide limitations

  • This path installs skills from a team library; administrators creating a library are directed to the separate desktop-app guide.
  • The README does not enumerate the other tools offered by the app.
Evidence and transparency

Sources & review

Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
FrancyJGLisboa/agent-skills-platform · Official homepage title · retained in its source language