OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

Agent Scripts

Agent Scripts is an MIT-licensed developer repository containing shared agent instructions, reusable skills, dependency-light shell helpers and local validation hooks for Codex- and Claude-style agents in local workspaces.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Agent Scripts — 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 agent instructions, skills and local helpers need to be maintained centrally and exposed across several repositories or agent environments.

Who it is for

  • Developers maintaining multiple local repositories.
  • Teams standardizing instructions and reusable workflows for Codex- or Claude-style agents.
  • Builders who need portable shell helpers or local skill validation.

Documented uses

  • Share agent rules across repositories through pointer-style instruction files.
  • Package reusable workflows in routable skill directories.
  • Synchronize skills into the different local layouts expected by Codex and Claude Code.
  • Validate skill metadata and maintain local workflow guardrails.
  • Provide small shell and TypeScript helpers for recurring repository tasks.
Before you choose

Requirements & limits

The repository is public and identifies an MIT license. Its documented use is local and self-hosted through CLI-oriented scripts.

  • The synchronization logic produces different layouts for Codex and Claude Code, including whole-root links for Codex and flat per-skill links for Claude Code.
  • Skills require YAML front matter with name and description; descriptions are intended for routing rather than extensive documentation.
  • The repository preserves real files and directories during ordinary synchronization and reports destination conflicts.
  • A local Git hook can use git config core.hooksPath hooks to enable the documented hook directory.
  • The browser helper is standalone and includes commands for navigation, JavaScript evaluation, screenshots, console and network inspection, and content search.

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

Add and validate a repository-local agent skill

Make a new skill available to the repository’s skill-routing structure and validate its metadata.

Before you start

  • A local checkout of Agent Scripts.
  • A new directory under skills/ containing a SKILL.md file.
  • YAML front matter with name and description fields.
  • The repository’s scripts/validate-skills helper.

Steps

  1. Create a new skill directory under skills/.

  2. Add SKILL.md with YAML front matter containing a name and short description.

  3. Keep the skill body terse and operational, placing repeatable commands in a scripts/ subdirectory when appropriate.

  4. Run scripts/validate-skills after editing.

Expected outcome

The skill’s documented structure is checked for YAML front matter and the required name and description fields.

Guide limitations

  • The source documents the file structure and validation step but does not demonstrate a complete runtime integration test.
  • The evidence does not specify the full execution environment for local hooks.
Documentation 1 ↗
Source-based practical guide

Synchronize skills for Codex and Claude Code

Build the documented local skill mirrors and shared instruction pointers after cloning or adding skills.

Before you start

  • A local checkout of Agent Scripts.
  • The documented local workspace layout, including the relevant repository paths.
  • Codex and/or Claude Code local skill locations.

Steps

  1. Run scripts/sync-skills after cloning or adding skills.

  2. Use the ordinary sync mode to build Codex whole-root links, Claude flat per-skill links and shared instruction pointers.

  3. For the documented legacy nested topology, invoke scripts/sync-skills --repair-nested-self-links with an explicit allowlist; add --dry-run for the scoped preview mode.

  4. Review reported collisions, stale links or refused changes.

Expected outcome

The documented Codex and Claude skill mirrors and shared instruction pointers are synchronized while real destination files and directories are preserved.

Guide limitations

  • Claude Code loads only one skill-directory level, so category subfolders are not scanned.
  • Name collisions resolve in the documented order and may be skipped.
  • Nested-link repair is not atomic concurrency protection; completed removals are not rolled back after a later error.
  • The source documents setup and validation behavior, not a complete agent workflow execution.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
2 October 2026
Original website title
steipete/agent-scripts · Official homepage title · retained in its source language
How we review sourcesSuggest a correction