Metadata-Version: 2.4
Name: workflow_auto_assembler
Version: 0.1.1
Summary: Experimental schema-first workflow synthesis tool for assembling simple typed workflows from existing functions.
Home-page: https://kiril-mordan.github.io/adaptive-reusables/workflow_auto_assembler/
Author: Kyrylo Mordan
Author-email: parachute.repo@gmail.com
License: apache-2.0
Keywords: aa-paa-tool
Classifier: Development Status :: 3 - Alpha
Description-Content-Type: text/markdown
License-File: LICENSE
License-File: NOTICE
Requires-Dist: pyyaml
Requires-Dist: attrs
Requires-Dist: attrsx
Requires-Dist: pydantic
Provides-Extra: openai
Requires-Dist: openai; extra == "openai"
Provides-Extra: ollama
Requires-Dist: ollama; extra == "ollama"
Provides-Extra: all
Requires-Dist: openai; extra == "all"
Requires-Dist: ollama; extra == "all"
Dynamic: author
Dynamic: author-email
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: keywords
Dynamic: license
Dynamic: license-file
Dynamic: provides-extra
Dynamic: requires-dist
Dynamic: summary

# Workflow auto assembler

Workflow Auto Assembler (WAA) is an experimental schema-first workflow synthesis tool.
It uses an LLM to assemble simple executable workflows from a task description, target
input/output models, and a catalog of available typed tools.

WAA is built around a narrow idea: instead of asking the model to solve a task
directly, expose tools through explicit schemas and let the model construct a linear
workflow whose intermediate states can be validated and tested. The resulting workflow
is persisted as an explicit artifact with I/O mappings, which makes it inspectable,
re-runnable, and easier to validate than a transient agent trace.

### Why Workflow-Auto-Assembler?

WAA treats tool selection and wiring as a schema-matching problem. When the required
capabilities exist in the available tool catalog, the assembled workflow can behave
like a larger typed function composed from smaller typed functions.

This package should be read as experimental adaptive tooling rather than a claim of
general autonomous planning. Its strongest use case is simple linear workflow assembly
under explicit schema constraints, with runner-based validation used to check that the
assembled workflow satisfies the requested output contract.

