JA-Agentic-Components

Unified repository for building and maintaining reusable components of the agentic stack: Skills, Custom Agents, Hooks, Prompts, and Orchestration assets.

Project Objective

This project serves as a single source of truth for reusable AI agent building blocks.

Core goals:

• Centralize Skills, Custom Agents, Hooks, Prompts, and Orchestration resources in one place.
• Make components easy to discover, reuse, and evolve across multiple AI coding environments.
• Document conventions and operational patterns so teams can compose reliable agentic systems.
• Support incremental growth from single skills to complete multi-agent workflows.

---

Tech Stack

• Language/runtime: Markdown-first repository (SKILL.md, README.md)
• Frameworks/standards: Agent Skill Standard (agentskills.io) and agent workflow conventions
• Tooling: Git, PowerShell (manual validation), VS Code/Copilot-compatible agent workflows

Prerequisites

• Git
• PowerShell (for manual validation commands)
• A compatible AI coding agent environment (for example GitHub Copilot, Claude Code, or Cursor)

Installation

git clone <your-fork-or-repo-url>
cd <repo-folder>

Install a single skill from a GitHub remote (npx)

Use this when you want to install only one skill from a remote skills repository:

Flag reference:

• -g: installs the skill globally (user-level, available across projects).
• -y: auto-confirms prompts (non-interactive mode).

Default (workspace/project-local) install:

npx skills add <owner>/<repo>@<skill> -y

Workspace note: omitting -g keeps the installation scoped to the current workspace/project.

Global install (user-level):

npx skills add <owner>/<repo>@<skill> -g -y

CLI note: npx -y ... is an npx flag, while ... -y at the end is parsed by skills add.

Workspace-local example:

npx skills add juanjoarranz/JA-Agentic-Components@project-memory -y

Global example:

npx skills add juanjoarranz/JA-Agentic-Components@project-memory -g -y

Components in this Repository

This repository includes reusable components across these domains:

• Skills
• Custom Agents
• Hooks
• Orchestration
• Prompts for software development

Additional resource:

• Developer Prompt Mastery: concise guide with practical prompt patterns to improve software development workflows.

Current implemented modules:

1) changelogger

• Path: .agents/skills/changelogger/SKILL.md
• Description: Automatically updates CHANGELOG.md with intended changes before commit finalization.
• Example usage prompts:
  • "Update the changelog before I commit these staged changes."
  • "Run changelogger with this conventional commit message."

2) commiter

• Path: .agents/skills/commiter/SKILL.md
• Description: Creates Git commits using Conventional Commits with detailed commit bodies.
• Example usage prompts:
  • "Commit my staged files using the commiter skill."
  • "Generate a conventional commit message with a detailed body for this change."

3) feature-docs

• Path: .agents/skills/feature-docs/SKILL.md
• Description: Generates single-file, verified feature documentation with code evidence for non-enterprise projects.
• Example usage prompts:
  • "Create quick feature docs for this module using feature-docs."
  • "Generate a verified single-file feature README with code references."

4) obsidian-plugin-helper

• Path: .agents/skills/obsidian-plugin-helper/SKILL.md
• Description: Provides comprehensive Obsidian plugin development guidance, including ESLint rules, TypeScript best practices, API usage (requestUrl vs fetch), UI/UX standards, and submission requirements.
• Example usage prompts:
  • "Help me implement this Obsidian plugin feature following best practices."
  • "Review this main.ts and manifest.json using the obsidian-plugin-helper guidance."

5) project-memory

• Path: .agents/skills/project-memory/SKILL.md
• Description: Sets up and maintains docs/project_notes/ to track bugs, decisions, key facts, and work history.
• Example usage prompts:
  • "Set up project memory for this repository."
  • "Log this bug fix and update our project memory notes."

6) readme-maintainer

• Path: .agents/skills/readme-maintainer/SKILL.md
• Description: Creates or updates the root README.md based on repository conventions and verified project context.
• Example usage prompts:
  • "Update the repository README to match the current project structure."
  • "Refresh project documentation in README format."

Project Structure

.
├── .github/
│   └── copilot-instructions.md
├── .agents/
│   └── skills/
│       ├── changelogger/
│       ├── commiter/
│       ├── feature-docs/
│       ├── obsidian-plugin-helper/
│       ├── project-memory/
│       └── readme-maintainer/
└── .vscode/

As the repository evolves, component directories for custom agents, hooks, and orchestration assets are added under this same project scope.

Development Conventions

• Skills live in .agents/skills/<skill-name>/.
• Skill required entry point: SKILL.md (uppercase).
• Skill names and directories use kebab-case.
• SKILL.md uses skill frontmatter with name and description.
• Keep instructions procedural and explicit, including output contracts.
• Apply the same clarity and reproducibility principles to custom agents, hooks, and orchestration definitions.

Troubleshooting

• Skill not detected by agents: verify SKILL.md exists at .agents/skills/<skill-name>/SKILL.md.
• Parsing issues: check frontmatter fence format and required fields (name, description).
• General component issues: validate paths, naming, and integration points against the corresponding component specification.

Contributing

1. Create or update a component (skill, custom agent, hook, or orchestration asset).
2. Follow the relevant naming and specification conventions.
3. Validate structure and references before opening a PR.
4. Open a PR with a clear summary of added/changed behavior.

License

This project is licensed under the Apache License 2.0.

See the full text in LICENSE.