Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/xvnpw/fabric-agent-action
🤖 A GitHub action that leverages fabric patterns through an agent-based approach
https://github.com/xvnpw/fabric-agent-action
agents ai fabric llm
Last synced: about 2 months ago
JSON representation
🤖 A GitHub action that leverages fabric patterns through an agent-based approach
- Host: GitHub
- URL: https://github.com/xvnpw/fabric-agent-action
- Owner: xvnpw
- License: mit
- Created: 2024-11-06T12:35:21.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2024-11-15T19:13:39.000Z (about 2 months ago)
- Last Synced: 2024-11-15T20:20:59.102Z (about 2 months ago)
- Topics: agents, ai, fabric, llm
- Language: Python
- Homepage:
- Size: 524 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Fabric Agent Action
[](https://github.com/xvnpw/fabric-agent-action/actions/workflows/ci.yaml)
🤖 **Fabric Agent Action** is a GitHub Action that leverages [Fabric Patterns](https://github.com/danielmiessler/fabric/tree/main/patterns) to automate complex workflows using an agent-based approach. Built with [LangGraph](https://www.langchain.com/langgraph), it intelligently selects and executes patterns using Large Language Models (LLMs).
## Features
- **Seamless Integration:** Easily incorporate the action into your existing workflows without additional setup.
- **Multi-Provider Support:** Choose between OpenAI, OpenRouter, or Anthropic based on your preference and availability.
- **Configurable Agent Behavior:** Select agent types (`router` or `react`) and customize their behavior to suit your workflow needs.
- **Flexible Pattern Management:** Include or exclude specific Fabric Patterns to optimize performance and comply with model limitations.
## Setup
Add the Fabric Agent Action to your workflow by referencing it in your `.yaml` file:
```yaml
- name: Execute Fabric Agent Action
uses: xvnpw/[email protected]
with:
input_file: path/to/input.md
output_file: path/to/output.md
env:
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
```
Set Environment Variables: Ensure you set the required API keys in your repository's secrets.
## Security
**⚠️ Important:** Before using this action, implement protections against unauthorized use, especially in public repositories. Unauthorized usage can lead to excessive API consumption and incur costs.
Use workflow conditions to limit who can run this action:
| Type | Abuse Description | Example Protecting Condition |
| --- | --- | --- |
| Pull request | Pull requests can originate from forks | `if: github.event.pull_request.head.repo.full_name == github.repository` |
| Issue comment | Anyone can create issues and add comments on public repositories | `github.event.comment.user.login == github.event.repository.owner.login` |
## Configuration
### Action Inputs
| Input | Description | Default |
|-------|-------------|---------|
| `input_file` | **Required** Source file containing input and agent instructions | |
| `output_file` | **Required** Destination file for pattern results | |
| `verbose` | Enable INFO level logging | `false` |
| `debug` | Enable DEBUG level logging | `false` |
| `agent_type` | Agent behavior model (`router`/`react`) | `router` |
| `agent_provider` | LLM provider for agent (`openai`/`openrouter`/`anthropic`) | `openai` |
| `agent_model` | Model name for agent | `gpt-4o` |
| `agent_temperature` | Model creativity (0-1) for agent | `0` |
| `agent_preamble_enabled` | Enable preamble in output | `false` |
| `agent_preamble` | Preamble added to the beginning of output | `##### (🤖 AI Generated)` |
| `fabric_provider` | Pattern execution LLM provider | `openai` |
| `fabric_model` | Pattern execution LLM model | `gpt-4o` |
| `fabric_temperature` | Pattern execution creativity (0-1) | `0` |
| `fabric-patterns-included` | Patterns to include (comma-separated). **Required for models with pattern limits (e.g., `gpt-4o`).** | |
| `fabric-patterns-excluded` | Patterns to exclude (comma-separated) | |
| `fabric_max_num_turns` | Maximum number of turns to LLM when running fabric patterns | 10 |
> **Note:** Models like `gpt-4o` have a limit on the number of tools (128), while Fabric currently includes 175 patterns (as of November 2024). Use `fabric_patterns_included` or `fabric_patterns_excluded` to tailor the patterns used. For access to all patterns without tool limits, consider using `claude-3-5-sonnet-20240620`.
Find the list of available Fabric Patterns [here](https://github.com/danielmiessler/fabric/tree/main/patterns).
### Required Environment Variables
Set one of the following API keys:
- `OPENAI_API_KEY`
- `OPENROUTER_API_KEY`
- `ANTHROPIC_API_KEY`
## Usage Example
This action is flexible in workflow integration and can be used on issues, pushes, etc.
### Issue Comments - Created or Edited
Below is an example of how to integrate the Fabric Agent Action into a GitHub Actions workflow that reacts to issue comments:
```yaml
name: Fabric Pattern Processing
on:
issue_comment:
types: [created, edited]
jobs:
process_fabric:
# Only trigger when comments start with '/fabric' from the repository owner
if: >
github.event.comment.user.login == github.event.repository.owner.login &&
startsWith(github.event.comment.body, '/fabric') &&
!github.event.issue.pull_request
runs-on: ubuntu-latest
permissions:
issues: write
contents: write
steps:
- name: Checkout
uses: actions/checkout@v4
# Use github-script to fetch issue and comment details and prepare input
- name: Prepare Input
uses: actions/github-script@v7
id: prepare-input
with:
script: |
const issue = await github.rest.issues.get({
issue_number: context.issue.number,
owner: context.repo.owner,
repo: context.repo.repo
});
const comment = await github.rest.issues.getComment({
comment_id: context.payload.comment.id,
owner: context.repo.owner,
repo: context.repo.repo
});
const input = `${comment.data.body}\n\n${issue.data.body}`;
require('fs').writeFileSync('fabric_input.md', input);
return input;
- name: Execute Fabric Patterns
uses: docker://ghcr.io/xvnpw/fabric-agent-action:v0.0.26
with:
input_file: fabric_input.md
output_file: fabric_output.md
agent_preamble_enabled: true
agent_model: gpt-4o # IMPORTANT: gpt-4o supports only 128 patterns; use fabric_patterns_included/fabric_patterns_excluded
fabric_patterns_included: clean_text,create_stride_threat_model,create_design_document,review_design,refine_design_document,create_threat_scenarios,improve_writing
env:
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
# Post the results back to the original issue
- name: Post Results
uses: peter-evans/create-or-update-comment@v4
with:
issue-number: ${{ github.event.issue.number }}
body-path: fabric_output.md
```
In this workflow:
- The job runs only when:
- The comment starts with `/fabric`.
- The comment author is the repository owner.
- The issue is not a pull request.
- This prevents unauthorized users from triggering the action, avoiding excessive API usage or costs.
### More Examples
| Example | Links |
| --- | --- |
| Create a pull request on changes in `README.md` to run the [improve_writing pattern](https://github.com/danielmiessler/fabric/blob/main/patterns/improve_writing/system.md) | [Pull request](https://github.com/xvnpw/fabric-agent-action-examples/pull/4), [workflow](https://github.com/xvnpw/fabric-agent-action-examples/blob/main/.github/workflows/fabric-readme-pr.yml) |
| Create a pull request on changes in the `docs/` directory to run the `improve_writing` pattern | [Pull request](https://github.com/xvnpw/fabric-agent-action-examples/pull/8), [workflow](https://github.com/xvnpw/fabric-agent-action-examples/blob/main/.github/workflows/fabric-docs-pr.yml) |
| Run fabric patterns from issue comments using the [router agent](#router-agent) | [Issue](https://github.com/xvnpw/fabric-agent-action-examples/issues/5), [workflow](https://github.com/xvnpw/fabric-agent-action-examples/blob/main/.github/workflows/fabric-issue-agent-router.yml) |
| Run fabric patterns from issue comments using the [react agent](#react-agent) | [Issue](https://github.com/xvnpw/fabric-agent-action-examples/issues/6), [workflow](https://github.com/xvnpw/fabric-agent-action-examples/blob/main/.github/workflows/fabric-issue-agent-react.yml) |
| Automatically run the fabric [write_pull_request pattern](https://github.com/danielmiessler/fabric/blob/main/patterns/write_pull-request/system.md) on pull requests | [Pull request](https://github.com/xvnpw/fabric-agent-action-examples/pull/7), [workflow](https://github.com/xvnpw/fabric-agent-action-examples/blob/main/.github/workflows/fabric-pr-diff.yml) |
## Agent Types
Agents select appropriate Fabric patterns based on the input. If no matching patterns exist, they return "no fabric pattern for this request" and terminate.
```mermaid
quadrantChart
title Agents: Autonomy vs. Reliability
x-axis Low Autonomy --> High Autonomy
y-axis Low Reliability --> High Reliability
Router Agent: [0.25, 0.75]
ReAct Agent: [0.6, 0.4]
```
In practice, there's often a trade-off between autonomy and reliability. Increasing LLM autonomy can sometimes reduce reliability due to factors like non-determinism or errors in tool selection.
### `router` Agent
Executes a single pattern selection with direct output:
```mermaid
%%{init: {'flowchart': {'curve': 'linear'}}}%%
graph TD;
__start__([
__start__
]):::first
assistant(assistant)
tools(tools)
__end__([__end__
]):::last
__start__ --> assistant;
tools --> __end__;
assistant -.-> tools;
assistant -.-> __end__;
classDef default fill:#f2f0ff,line-height:1.2
classDef first fill-opacity:0
classDef last fill:#bfb6fc
```
**Example Input:**
```markdown
/fabric improve writing
I encountered a challenge in creating high-quality design documents for my threat modeling research. About a year and a half ago, I created AI Nutrition-Pro architecture and have been using it since then. What if it's already in LLMs' training data? Testing threat modeling capabilities could give me false results.
```
**Example Output:**
```markdown
##### (🤖 AI Generated)
I encountered a challenge in creating high-quality design documents for my threat modeling research. About a year and a half ago, I developed the AI Nutrition-Pro architecture and have been using it since then. What if it's already included in the training data of LLMs? Testing threat modeling capabilities could yield false results.
```
### `react` Agent
The agent takes input from the user, decides on pattern selection, and reasons about the output from the pattern:
```mermaid
%%{init: {'flowchart': {'curve': 'linear'}}}%%
graph TD;
__start__([
__start__
]):::first
assistant(assistant)
tools(tools)
__end__([__end__
]):::last
__start__ --> assistant;
tools --> assistant;
assistant -.-> tools;
assistant -.-> __end__;
classDef default fill:#f2f0ff,line-height:1.2
classDef first fill-opacity:0
classDef last fill:#bfb6fc
```
Use `fabric_max_num_turns` to limit number of turns to LLM for pattern processing (=tools node).
This approach follows the intuition behind [ReAct](https://react-lm.github.io/):
- **Act:** Let the model call specific tools (patterns).
- **Observe:** Pass the tool output back to the model.
- **Reason:** Let the model reason about the tool output to decide the next action.
**Example Input:**
```markdown
/fabric clean text and improve writing
I encountered a challenge in creating high-quality design documents for my threat modeling research. About a year and a half ago, I created AI Nutrition-Pro architecture and have been using it since then. What if it's already in LLMs' training data? Testing threat modeling capabilities could give me false results.
```
**Example Output:**
```markdown
##### (🤖 AI Generated)
Here is the cleaned and improved version of your text:
Cleaned Text:
"I encountered a challenge in creating high-quality design documents for my threat modeling research. About a year and a half ago, I created AI Nutrition-Pro architecture and have been using it since then. What if it's already in LLMs' training data? Testing threat modeling capabilities could give me false results."
Improved Writing:
"I encountered a challenge in creating high-quality design documents for my threat modeling research. About a year and a half ago, I developed the AI Nutrition-Pro architecture and have been using it since then. What if it's already included in the training data of LLMs? Testing threat modeling capabilities could yield false results."
```
## Debugging
You have two ways to gain insights into the internal workings of the system.
### Debug Argument
Enable the `debug` mode by setting the `debug` argument to `true`. This will provide more detailed information about the process.
### LangSmith
[LangSmith](https://www.langchain.com/langsmith) offers a free tier that lets you visually track interactions between agents and LLMs, making debugging easier.
To use LangSmith, set the following environment variables:
- `LANGCHAIN_API_KEY`: `${{ secrets.LANGCHAIN_API_KEY }}`
- `LANGCHAIN_TRACING_V2`: `true`
## Supported LLM Providers
- [OpenAI](https://platform.openai.com/) - Industry standard
- [OpenRouter](https://openrouter.ai/) - Multi-model gateway
- [Anthropic](https://www.anthropic.com/) - Claude models
## Contributing
Contributions are welcome! Please open issues and pull requests. Ensure that you follow the existing code style and include tests for new features.
## License
This project is licensed under the [MIT License](LICENSE).