mirror of
https://github.com/FuzzingLabs/fuzzforge_ai.git
synced 2026-02-12 20:32:46 +00:00
d786c6dab1
* fix: block Podman on macOS and remove ghcr.io default - Add platform check in PodmanCLI.__init__() that raises FuzzForgeError on macOS with instructions to use Docker instead - Change RegistrySettings.url default from "ghcr.io/fuzzinglabs" to "" (empty string) for local-only mode since no images are published yet - Update _ensure_module_image() to show helpful error when image not found locally and no registry configured - Update tests to mock Linux platform for Podman tests - Add root ruff.toml to fix broken configuration in fuzzforge-runner * rewrite guides for module architecture and update repo links --------- Co-authored-by: AFredefon <antoinefredefon@yahoo.fr>
FuzzForge Runner
Direct execution engine for FuzzForge OSS. Provides simplified module and workflow execution without requiring Temporal or external infrastructure.
Overview
The Runner is designed for local-first operation, executing FuzzForge modules directly in containerized sandboxes (Docker/Podman) without workflow orchestration overhead.
Features
- Direct module execution in isolated containers
- Sequential workflow orchestration (no Temporal required)
- Local filesystem storage (S3 optional)
- SQLite-based state management (optional)
Usage
from fuzzforge_runner import Runner
from fuzzforge_runner.settings import Settings
settings = Settings()
runner = Runner(settings)
# Execute a single module
result = await runner.execute_module(
module_identifier="my-module",
project_path="/path/to/project",
)
# Execute a workflow (sequential steps)
result = await runner.execute_workflow(
workflow_definition=workflow,
project_path="/path/to/project",
)
Configuration
Environment variables:
FUZZFORGE_STORAGE_PATH: Local storage directory (default:~/.fuzzforge/storage)FUZZFORGE_ENGINE_TYPE: Container engine (dockerorpodman)FUZZFORGE_ENGINE_SOCKET: Container socket path