---
# SkillLink
**Updates:**
**Update (July 2025):**
* **SkillLink has been reimagined and is now a major upgrade of the former SkillsManager package.
* **The original SkillLink package (which handled syncing skills/files) has been renamed and upgraded to SyncLink.
* **If you need advanced skill management and execution with LLMs, use SkillLink.
* **If you need to sync files or skills between local and remote, use SyncLink instead.
* **The SkillsManager package is now deprecated and will receive no further updates. Please migrate to SkillLink for all future development.
**Migration Tips:**
* **If you previously used SkillsManager, upgrade to SkillLink—it's fully compatible and adds new features.
* **If you previously used SkillLink for syncing, switch to SyncLink for the latest version and improvements.
* **[SyncLink](https://pypi.org/project/SyncLink/)**.
## Overview
**SkillLink** is a next-generation framework for managing and executing skills/actions with large language models (**LLMs**) using **natural language**—not vendor-bound JSON or TYPED schemas definitions.
Unlike traditional function-calling, SkillLink lets your LLM output actions in freeform(Natural Language) text. The SkillLink framework interprets, maps, and executes those actions using your own Python code—no matter what you call your skills/actions or where you store them.
**Highlights:**
* **Zero vendor lock-in:** Not tied to any model provider’s format or schema.
* **No naming restrictions:** Name your skills, folders, or groups anything you like—SkillLink doesn’t care.
* **No JSON schemas:** No decorators, no rigid typing, no endless schema maintenance.
* **Natural language as the interface:** The LLM outputs actions in text, not function\_call JSON or TYPED.
* **Add/organize skills freely:** Just point SkillLink at your code—group, reload, or restrict however you want.
* **Centralized execution:** All skill mapping, argument parsing, and execution happens in one place.
**Backward compatible:**
* **Want to stick with traditional function calling?** No problem. SkillLink can still work with JSON and TYPED schemas definitions if you prefer that style.
* **Want to use SkillLink with existing function calls?** You can still use it alongside traditional function calling, so you can migrate gradually or keep both styles.
> **NOTE:**
> If you want to use JSON schemas or TYPED definitions (like OpenAI, Gemini, Anthropic, etc.),
> **SkillLink automatically generates and manages all required schemas for you based on your code and docstrings—**
> **You NEVER have to write, maintain, or register schemas manually.**
---
## Why SkillLink?
Traditional LLM “function calling” means:
* Rigid function and argument definitions.
* Tedious JSON and TYPED schema management.
* Forced adherence to provider conventions.
* Locked into one vendor’s API.
**SkillLink:**
* Uses only natural language—models describe actions however they want.
* Lets you migrate skills and switch model providers instantly.
* Removes all JSON and TYPED schema headaches.
* Lets you organize, group, and name skills however you want.
---
## Key Features
* **Flexible Skill Loading:**
Add, reload, or restrict any set of skills, with any folder or group name.
* **Natural Language Action Parsing:**
Models output plain English (or any language), and SkillLink interprets it to execute your code.
* **Centralized Execution:**
No scattered registries—skill mapping, argument parsing, and execution are all in one place.
---
## Organizing and Naming Your Skills
There are **no reserved names, no required folder layout, and no fixed naming conventions** in SkillLink.
You control all naming, grouping, and structure—use whatever makes sense for your team, project, or workflow.
**Examples:**
* **Refer to SMExamples Directory or the Examples on GitHub:**
**Just point SkillLink at any directory or skill group you want—no restrictions, ever.**
---
### Example Layouts (ALL are valid)
```
project_root/
├── Skills/ # This is just an example, you can name it whatever you want, you can even nest the directories it does not have to be at the root
│ ├── Foo/ # Can do nested folders, any names
│ │ ├── summarize.py
│ │ └── summarize_pdf.py
│ ├── Bar/
│ │ ├── sendEmail.py
│ │ └── joke_skill.py
│ ├── EmailTool.py # Or directly in the directory
│ └── ping_tool.py
├── Tools/ # This is just an example, you can name it whatever you want, you can even nest the directories it does not have to be at the root
│ ├── Foo/ # Can do nested folders, any names
│ │ ├── summarize.py
│ │ └── summarize_pdf.py
│ ├── Bar/
│ │ ├── sendEmail.py
│ │ └── joke_skill.py
│ ├── EmailTool.py # Or directly in the directory
│ └── ping_tool.py
├── SkillLoader.py # Whatever you want to call it
├── .env
└── ...
```
Or any structure you prefer.
---
## How It Works
1. **Model outputs natural language:**
"Summarize the latest report and update our dashboard."
2. **SkillLink parses the text** and finds the right skills (whatever you’ve named them).
3. **Skills are executed** and results returned—no schemas, no JSON or TYPED schemas, just Python code.
---
## Example: Comparing Approaches
### 1. With SkillLink (No Naming Restrictions)
**Behind the scenes:**
* Model outputs: “Summarize PDF, then send summary via email.”
* SkillLink parses and executes your skills, regardless of how you’ve named or grouped them.
* No decorators, no JSON no TYPED schemas.
---
### 2. Traditional Function/Tool Calling (Provider Schema Required)
* Must define every function and argument as a JSON or TYPED schema.
* Register all with the provider.
* Model outputs strict `function_call` blobs.
* You parse, validate, and execute under the provider’s constraints.
> **SkillLink does this part for you automatically if you ever need it—no manual schemas, no boilerplate.**
---
## Adding and Organizing Skills
* Drop Python modules into any folder, with any name or grouping.
* Point SkillLink at those folders in your class (`loadComponents`).
* No JSON or TYPED, no schemas, no required naming—total freedom.
---
## Why Use SkillLink?
* **Never locked into a provider.**
* **No JSON or TYPED schema maintenance.**
* **Flexible skill organization:** Name and group skills however you want.
* **Production ready:** Clean, scalable, and easy to extend.
---
## FAQ
**Q: Do I need to follow a specific naming or folder structure?**
A: No. You can use any names and any folder hierarchy.
**Q: Can I use this with any LLM PROVIDER?**
A: Yes.
**Q: How does SkillLink know what to execute?**
A: It parses the model’s plain language output and matches actions to your Python skills, regardless of naming.
**Q: Do I have to write JSON or TYPED schemas or tool definitions?**
A: **Never. SkillLink auto-generates and manages them for you when required.**
---
## Code Examples
You can find code examples on my [GitHub repository](https://github.com/TristanMcBrideSr/TechBook).
---
## License
This project is licensed under the [Apache License, Version 2.0](LICENSE).
Copyright 2025 Tristan McBride Sr.
---
## Acknowledgements
Project by:
- Tristan McBride Sr.
- Sybil
Raw data
{
"_id": null,
"home_page": null,
"name": "SkillLink",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": "\"Tristan McBride Sr.\" <142635792+TristanMcBrideSr@users.noreply.github.com>",
"keywords": "AI, Agents, Skills, Tools, Autoload, Productivity, Automation",
"author": "Tristan McBride Sr.",
"author_email": "\"Tristan McBride Sr.\" <142635792+TristanMcBrideSr@users.noreply.github.com>",
"download_url": "https://files.pythonhosted.org/packages/8a/77/b0b938ffb174a09bf4ccfbacc4391706689b784ad2e093fcfb85a1b1b2e1/skilllink-0.2.7.tar.gz",
"platform": null,
"description": "\ufeff\r\n---\r\n\r\n# SkillLink\r\n\r\n**Updates:**\r\n\r\n**Update (July 2025):**\r\n\r\n* **SkillLink has been reimagined and is now a major upgrade of the former SkillsManager package.\r\n\r\n* **The original SkillLink package (which handled syncing skills/files) has been renamed and upgraded to SyncLink.\r\n\r\n* **If you need advanced skill management and execution with LLMs, use SkillLink.\r\n\r\n* **If you need to sync files or skills between local and remote, use SyncLink instead.\r\n\r\n* **The SkillsManager package is now deprecated and will receive no further updates. Please migrate to SkillLink for all future development.\r\n\r\n**Migration Tips:**\r\n\r\n* **If you previously used SkillsManager, upgrade to SkillLink\u2014it's fully compatible and adds new features.\r\n\r\n* **If you previously used SkillLink for syncing, switch to SyncLink for the latest version and improvements.\r\n* **[SyncLink](https://pypi.org/project/SyncLink/)**.\r\n\r\n## Overview\r\n\r\n**SkillLink** is a next-generation framework for managing and executing skills/actions with large language models (**LLMs**) using **natural language**\u2014not vendor-bound JSON or TYPED schemas definitions.\r\nUnlike traditional function-calling, SkillLink lets your LLM output actions in freeform(Natural Language) text. The SkillLink framework interprets, maps, and executes those actions using your own Python code\u2014no matter what you call your skills/actions or where you store them.\r\n\r\n**Highlights:**\r\n\r\n* **Zero vendor lock-in:** Not tied to any model provider\u2019s format or schema.\r\n* **No naming restrictions:** Name your skills, folders, or groups anything you like\u2014SkillLink doesn\u2019t care.\r\n* **No JSON schemas:** No decorators, no rigid typing, no endless schema maintenance.\r\n* **Natural language as the interface:** The LLM outputs actions in text, not function\\_call JSON or TYPED.\r\n* **Add/organize skills freely:** Just point SkillLink at your code\u2014group, reload, or restrict however you want.\r\n* **Centralized execution:** All skill mapping, argument parsing, and execution happens in one place.\r\n\r\n**Backward compatible:**\r\n\r\n* **Want to stick with traditional function calling?** No problem. SkillLink can still work with JSON and TYPED schemas definitions if you prefer that style.\r\n* **Want to use SkillLink with existing function calls?** You can still use it alongside traditional function calling, so you can migrate gradually or keep both styles.\r\n\r\n> **NOTE:**\r\n> If you want to use JSON schemas or TYPED definitions (like OpenAI, Gemini, Anthropic, etc.),\r\n> **SkillLink automatically generates and manages all required schemas for you based on your code and docstrings\u2014**\r\n> **You NEVER have to write, maintain, or register schemas manually.**\r\n\r\n---\r\n\r\n## Why SkillLink?\r\n\r\nTraditional LLM \u201cfunction calling\u201d means:\r\n\r\n* Rigid function and argument definitions.\r\n* Tedious JSON and TYPED schema management.\r\n* Forced adherence to provider conventions.\r\n* Locked into one vendor\u2019s API.\r\n\r\n**SkillLink:**\r\n\r\n* Uses only natural language\u2014models describe actions however they want.\r\n* Lets you migrate skills and switch model providers instantly.\r\n* Removes all JSON and TYPED schema headaches.\r\n* Lets you organize, group, and name skills however you want.\r\n\r\n---\r\n\r\n## Key Features\r\n\r\n* **Flexible Skill Loading:**\r\n Add, reload, or restrict any set of skills, with any folder or group name.\r\n\r\n* **Natural Language Action Parsing:**\r\n Models output plain English (or any language), and SkillLink interprets it to execute your code.\r\n\r\n* **Centralized Execution:**\r\n No scattered registries\u2014skill mapping, argument parsing, and execution are all in one place.\r\n\r\n---\r\n\r\n## Organizing and Naming Your Skills\r\n\r\nThere are **no reserved names, no required folder layout, and no fixed naming conventions** in SkillLink.\r\nYou control all naming, grouping, and structure\u2014use whatever makes sense for your team, project, or workflow.\r\n\r\n**Examples:**\r\n\r\n* **Refer to SMExamples Directory or the Examples on GitHub:**\r\n\r\n**Just point SkillLink at any directory or skill group you want\u2014no restrictions, ever.**\r\n\r\n---\r\n\r\n### Example Layouts (ALL are valid)\r\n\r\n```\r\nproject_root/\r\n\u251c\u2500\u2500 Skills/ # This is just an example, you can name it whatever you want, you can even nest the directories it does not have to be at the root\r\n\u2502 \u251c\u2500\u2500 Foo/ # Can do nested folders, any names\r\n\u2502 \u2502 \u251c\u2500\u2500 summarize.py\r\n\u2502 \u2502 \u2514\u2500\u2500 summarize_pdf.py\r\n\u2502 \u251c\u2500\u2500 Bar/\r\n\u2502 \u2502 \u251c\u2500\u2500 sendEmail.py\r\n\u2502 \u2502 \u2514\u2500\u2500 joke_skill.py\r\n\u2502 \u251c\u2500\u2500 EmailTool.py # Or directly in the directory\r\n\u2502 \u2514\u2500\u2500 ping_tool.py\r\n\u251c\u2500\u2500 Tools/ # This is just an example, you can name it whatever you want, you can even nest the directories it does not have to be at the root\r\n\u2502 \u251c\u2500\u2500 Foo/ # Can do nested folders, any names\r\n\u2502 \u2502 \u251c\u2500\u2500 summarize.py\r\n\u2502 \u2502 \u2514\u2500\u2500 summarize_pdf.py\r\n\u2502 \u251c\u2500\u2500 Bar/\r\n\u2502 \u2502 \u251c\u2500\u2500 sendEmail.py\r\n\u2502 \u2502 \u2514\u2500\u2500 joke_skill.py\r\n\u2502 \u251c\u2500\u2500 EmailTool.py # Or directly in the directory\r\n\u2502 \u2514\u2500\u2500 ping_tool.py\r\n\u251c\u2500\u2500 SkillLoader.py # Whatever you want to call it\r\n\u251c\u2500\u2500 .env\r\n\u2514\u2500\u2500 ...\r\n\r\n```\r\n\r\nOr any structure you prefer.\r\n\r\n---\r\n\r\n## How It Works\r\n\r\n1. **Model outputs natural language:**\r\n \"Summarize the latest report and update our dashboard.\"\r\n2. **SkillLink parses the text** and finds the right skills (whatever you\u2019ve named them).\r\n3. **Skills are executed** and results returned\u2014no schemas, no JSON or TYPED schemas, just Python code.\r\n\r\n---\r\n\r\n## Example: Comparing Approaches\r\n\r\n### 1. With SkillLink (No Naming Restrictions)\r\n\r\n**Behind the scenes:**\r\n\r\n* Model outputs: \u201cSummarize PDF, then send summary via email.\u201d\r\n* SkillLink parses and executes your skills, regardless of how you\u2019ve named or grouped them.\r\n* No decorators, no JSON no TYPED schemas.\r\n\r\n---\r\n\r\n### 2. Traditional Function/Tool Calling (Provider Schema Required)\r\n\r\n* Must define every function and argument as a JSON or TYPED schema.\r\n* Register all with the provider.\r\n* Model outputs strict `function_call` blobs.\r\n* You parse, validate, and execute under the provider\u2019s constraints.\r\n\r\n> **SkillLink does this part for you automatically if you ever need it\u2014no manual schemas, no boilerplate.**\r\n\r\n---\r\n\r\n## Adding and Organizing Skills\r\n\r\n* Drop Python modules into any folder, with any name or grouping.\r\n* Point SkillLink at those folders in your class (`loadComponents`).\r\n* No JSON or TYPED, no schemas, no required naming\u2014total freedom.\r\n\r\n---\r\n\r\n## Why Use SkillLink?\r\n\r\n* **Never locked into a provider.**\r\n* **No JSON or TYPED schema maintenance.**\r\n* **Flexible skill organization:** Name and group skills however you want.\r\n* **Production ready:** Clean, scalable, and easy to extend.\r\n\r\n---\r\n\r\n## FAQ\r\n\r\n**Q: Do I need to follow a specific naming or folder structure?**\r\nA: No. You can use any names and any folder hierarchy.\r\n\r\n**Q: Can I use this with any LLM PROVIDER?**\r\nA: Yes.\r\n\r\n**Q: How does SkillLink know what to execute?**\r\nA: It parses the model\u2019s plain language output and matches actions to your Python skills, regardless of naming.\r\n\r\n**Q: Do I have to write JSON or TYPED schemas or tool definitions?**\r\nA: **Never. SkillLink auto-generates and manages them for you when required.**\r\n\r\n---\r\n\r\n## Code Examples\r\n\r\nYou can find code examples on my [GitHub repository](https://github.com/TristanMcBrideSr/TechBook).\r\n\r\n---\r\n\r\n## License\r\n\r\nThis project is licensed under the [Apache License, Version 2.0](LICENSE).\r\nCopyright 2025 Tristan McBride Sr.\r\n\r\n---\r\n\r\n## Acknowledgements\r\n\r\nProject by:\r\n- Tristan McBride Sr.\r\n- Sybil\r\n",
"bugtrack_url": null,
"license": null,
"summary": "Modern manager for AI skills, tools, loading, execution, and automation.",
"version": "0.2.7",
"project_urls": {
"Homepage": "https://github.com/TristanMcBrideSr"
},
"split_keywords": [
"ai",
" agents",
" skills",
" tools",
" autoload",
" productivity",
" automation"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "95972f89fc1e0c2d0998d86a9348e5889fd6865866b1fda728c7e012b260ea42",
"md5": "317001981683c0ba948c0d00eeaf7fe6",
"sha256": "3271dba9435506b1cabaaa2d7e97c822586b624604cb160a7fcafa559d76785c"
},
"downloads": -1,
"filename": "skilllink-0.2.7-py3-none-any.whl",
"has_sig": false,
"md5_digest": "317001981683c0ba948c0d00eeaf7fe6",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 33542,
"upload_time": "2025-08-01T00:42:58",
"upload_time_iso_8601": "2025-08-01T00:42:58.084470Z",
"url": "https://files.pythonhosted.org/packages/95/97/2f89fc1e0c2d0998d86a9348e5889fd6865866b1fda728c7e012b260ea42/skilllink-0.2.7-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "8a77b0b938ffb174a09bf4ccfbacc4391706689b784ad2e093fcfb85a1b1b2e1",
"md5": "5da3585afaf9087d9bf330f80377458b",
"sha256": "2e8123f4e5415844a3819e3ac8d1227c04754b160e2feb9b43af32ad261120c2"
},
"downloads": -1,
"filename": "skilllink-0.2.7.tar.gz",
"has_sig": false,
"md5_digest": "5da3585afaf9087d9bf330f80377458b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 32680,
"upload_time": "2025-08-01T00:42:58",
"upload_time_iso_8601": "2025-08-01T00:42:58.955030Z",
"url": "https://files.pythonhosted.org/packages/8a/77/b0b938ffb174a09bf4ccfbacc4391706689b784ad2e093fcfb85a1b1b2e1/skilllink-0.2.7.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-01 00:42:58",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "skilllink"
}
JSON