skild

🥈Silver

Skild is a package manager for AI Agent Skills, enabling developers to discover, install, manage, and publish skills for AI agents. It streamlines the integration of AI capabilities into workflows, benefiting operations teams by reducing setup time and improving agent functionality.

10570Updated 3w ago

Intermediate30min to implementautomation

Saves ~105 min per use

Quick InstallView Source

git clone https://github.com/Peiiii/skild.git

Works with:

Claude

Overview

About This Skill

How to Use

[{"step":1,"action":"Identify the skill you need","details":"Browse the skild registry (skild search) or GitHub for skills matching your use case. Note the skill name, version, and source (official registry or custom repo).","tips":["Use `skild search [keyword]` to find skills by functionality (e.g., 'data_ingestion' or 'ticket_routing').","Check the skill’s README for compatibility notes, dependencies, and example configurations."]},{"step":2,"action":"Install or manage the skill","details":"Run the appropriate skild command to install (`skild install`), update (`skild update`), or remove (`skild uninstall`) the skill. Specify version or source if needed.","tips":["For custom skills, use `--source https://github.com/[user]/[repo]` to install from a GitHub repository.","Add `--dry-run` to preview changes before applying them."]},{"step":3,"action":"Configure the skill","details":"Locate the skill’s configuration file (usually in ~/.skild/config.yaml or a skill-specific directory) and set required parameters like API keys, model paths, or environment variables.","tips":["Use `skild config list` to view all configurable parameters for the installed skill.","For skills requiring models, download pre-trained models from the skill’s releases page or Hugging Face."]},{"step":4,"action":"Test the skill","details":"Run a test to validate the skill’s functionality with your data. Use `skild test [skill_name]` with input/output flags to process sample files.","tips":["Start with small datasets to verify outputs before scaling up.","Use `--verbose` to debug issues during testing."]},{"step":5,"action":"Integrate into your workflow","details":"Embed the skill into your AI agent’s pipeline by calling it via its API endpoint or SDK. Monitor performance and update configurations as needed.","tips":["For production use, set up logging (`skild logs [skill_name]`) to track errors and performance metrics.","Schedule periodic updates with `skild update` to ensure you’re using the latest versions."]}]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/Peiiii/skild

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use skild to [ACTION: discover/install/manage/publish] the AI skill '[SKILL_NAME]' for [PURPOSE: automation/testing/integration]. If installing, specify the version or source (e.g., 'from the official registry' or 'from GitHub repo [URL]'). Include any required dependencies or configuration flags.

Example Output

```json
{
  "action": "install",
  "skill_name": "data_ingestion",
  "version": "1.2.0",
  "source": "official registry",
  "dependencies": ["pandas>=2.0.0", "requests>=2.31.0"],
  "output": {
    "status": "success",
    "installed_path": "/usr/local/lib/skild/skills/data_ingestion",
    "dependencies_installed": true,
    "next_steps": [
      "Verify installation with: skild list",
      "Configure API keys in ~/.skild/config.yaml",
      "Test with: skild test data_ingestion --input sample.csv"
    ]
  }
}
```

**Example Workflow:**
You’re building an AI agent to automate customer support ticket routing. After researching skills in the skild registry, you decide to install `ticket_classifier` to categorize incoming tickets by urgency and topic. Using skild, you run:
```bash
skild install ticket_classifier --version 0.9.5 --source official_registry
```

The output confirms the skill is installed in `/opt/ai_agents/skills/ticket_classifier` and lists dependencies like `scikit-learn` and `spaCy`. A warning appears: "Missing config: Add `MODEL_PATH=/models/ticket_classifier_v0.9.5` to ~/.skild/config.yaml." You create the directory, download the pre-trained model from the skill’s GitHub releases, and update the config. Finally, you test the skill:
```bash
skild test ticket_classifier --input tickets.jsonl --output routed_tickets.json
```

The test processes 500 tickets, achieving 89% accuracy on your validation set. The routed_tickets.json file shows tickets labeled with `priority: high`, `topic: billing`, and `assigned_agent: alice@company.com`. You’re now ready to integrate the skill into your agent’s pipeline.