5D for PydanticAI approval-required toolsets

Pain it solves

PydanticAI gives you deferred tools and approvals, but you still need a policy that decides which tool calls should require approval and which ones should just run and log.

Best when you already use deferred tools and want runtime policy instead of hard-coded lists of actions that always require approval.

When to use 5D

Use 5D when your agent can write files, run shell, call external APIs, or touch sensitive tools.

In this setup, 5D returns a normalized runtime decision: allow, review, or deny, plus a tripwire_triggered flag for runtimes that want a simpler guardrail signal.

Install

git clone https://github.com/theDoc001/fivedrisk.git
cd fivedrisk
python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

Minimal example

from fivedrisk import DecisionLog, evaluate_action, load_policy
from pydantic_ai.toolsets import FunctionToolset

policy = load_policy("policy.yaml")
log = DecisionLog("fivedrisk.db")
toolset = FunctionToolset()

guarded = toolset.approval_required(
    lambda ctx, tool_def, tool_args: evaluate_action(
        tool_name=tool_def.name,
        tool_input=tool_args,
        policy=policy,
        log=log,
        source="pydanticai",
    ).requires_review
)