Assay Python Bindings

assay-it is the Python package for the assay import namespace.

It currently provides a small, bounded surface on top of Assay core:

• AssayClient for writing JSONL trace records from Python
• Coverage for policy coverage analysis against trace/tool-call inputs
• Explainer for rule-by-rule explanation of a single trace
• a pytest fixture plugin exposed as assay_client

Install

pip install assay-it

CPython 3.12, 3.13, and 3.14 on macOS x86_64/arm64 and Linux x86_64; other interpreters and platforms are not claimed.

Import path:

from assay import AssayClient, Coverage, Explainer, validate

What It Is

This package is for Python-side trace capture and lightweight policy-analysis helpers. It is useful when you want to:

• append structured trace events to a JSONL file from Python
• analyze tool-call traces against an Assay policy
• explain why a given trace matched or violated policy rules
• use a simple pytest fixture around the Python client

What It Is Not

This package is not the full Assay CLI or the full trust-compiler surface. It does not replace:

• assay mcp wrap and runtime MCP enforcement
• evidence bundle generation and verification workflows
• Trust Basis / Trust Card generation
• the GitHub Action or release-binary install path

Those outward-facing product surfaces live in the main Assay CLI and repository docs.

Example

from assay import AssayClient, Coverage, Explainer

client = AssayClient("traces.jsonl")
client.record_trace({"tool": "ToolA", "args": {"q": "weather"}})

coverage = Coverage("assay.yaml")
report = coverage.analyze([[{"tool": "ToolA", "args": {"q": "weather"}}]])

explainer = Explainer("assay.yaml")
details = explainer.explain([{"tool": "ToolA", "args": {"q": "weather"}}])

Repository

• Main project: https://github.com/Rul1an/assay
• Current CLI and trust-compiler docs: https://github.com/Rul1an/assay/blob/main/README.md