foxy-audit 1.0.0

Governance-as-Code SDK for AI compliance: local SHA-256 hashing + tamper-evident audit telemetry, with zero raw-text exfiltration.

foxy-audit (SDK)

Governance-as-Code for AI. One decorator → a tamper-evident, privacy-preserving audit trail.

The SDK hashes your LLM prompt + response locally (SHA-256), throws the raw text away, and streams only metadata to the Foxy Audit backend. It also fires a best-effort local UDP ping so the desktop "fox" companion reacts in real time (green on every logged call, red on a policy breach).

Install

pip install -e .            # from this sdk/ folder, for local development

Runtime dependency: requests only.

Use

import os
from foxy_audit import FoxyClient

foxy = FoxyClient(api_key=os.getenv("FOXY_API_KEY"))   # or just rely on the env var

@foxy.audit(policy="hipaa_basic")
def ask_model(prompt: str) -> str:
    return llm_client.generate(prompt)     # your existing code — unchanged

Every call to ask_model is now hashed, logged, and graded. Or use the module-level decorator, which builds a client from the environment:

from foxy_audit import audit

@audit(policy="soc2")
def summarize(text: str) -> str:
    ...

Attributing the model (agent)

Pass agent= to record which model produced the interaction. The backend folds it into the tamper-evident hash chain, so the attribution can't be altered after the fact:

@foxy.audit(policy="soc2", agent="gpt-4o")
def ask_model(prompt: str) -> str:
    ...

agent is optional — rows logged without it hash exactly as before, so existing chains keep verifying.

Configuration

Setting	Kwarg	Env var	Default
API key	api_key	FOXY_API_KEY	(none → HTTP disabled)
Backend URL	endpoint	FOXY_BACKEND_URL	http://127.0.0.1:8000
Desktop ping	desktop_ping	—	True (127.0.0.1:9999)

With no API key the SDK is a graceful no-op for the cloud path: it still runs your function and still pings the desktop fox, but skips the HTTP upload — so you can see the fox react before any backend exists.

Guarantees

• Never blocks your function — the HTTP upload runs on a background daemon thread.
• Never raises its own errors into your app — all telemetry failures are swallowed.
• Never sees raw text server-side — only SHA-256 digests + token count + policy tag leave the host.
• Works with both sync and async functions; the return value is always passed through unchanged.

What gets sent

To the backend (POST /v1/logs, Authorization: Bearer <key>):

{"prompt_hash": "<64 hex>", "response_hash": "<64 hex>", "token_count": 123, "policy_tag": "hipaa_basic"}

To the desktop fox (UDP 127.0.0.1:9999):

{"event": "hash_ok", "policy": "hipaa_basic", "tokens": 123, "ts": 1719300000}
{"event": "policy_breach", "reason": "...", "risk_score": 87, "policy": "hipaa_basic", "ts": ...}