decionis-autogen
Add a Decionis policy decision immediately before an AutoGen tool executes. The package supports
the current AutoGen Core tool interface and the legacy AutoGen 0.2 UserProxyAgent execution hook.
Current AutoGen
pip install decionis-autogen
from autogen_core.tools import FunctionTool
from decionis import DecionisClient
from decionis_autogen import guard_tool
refund_tool = FunctionTool(send_refund, description="Issue a customer refund.")
guarded_refund = guard_tool(
inner_tool=refund_tool,
client=DecionisClient(api_key="..."),
tenant_id="your-org-uuid",
workflow_key="refund_execution",
site_base_url="https://decionis.com",
)
agent = AssistantAgent(..., tools=[guarded_refund])
The wrapper preserves the inner tool schema. An ALLOW verdict delegates to the tool. BLOCK,
REVIEW_REQUIRED, and ESCALATE raise DecionisToolRefusal before the side effect runs.
AutoGen 0.2 UserProxyAgent
The requested execute_function interception applies to AutoGen 0.2. Install the legacy extra on
Python 3.10–3.12:
pip install 'decionis-autogen[legacy]'
from decionis import DecionisClient
from decionis_autogen import DecionisUserProxyAgent
executor = DecionisUserProxyAgent(
"executor",
decionis_client=DecionisClient(api_key="..."),
tenant_id="your-org-uuid",
function_map={"send_refund": send_refund},
human_input_mode="NEVER",
)
The subclass overrides both execute_function and a_execute_function. A blocked call returns a
structured function error with the dossier id and never invokes the registered function.
For either API, set shadow_mode=True to record would-block verdicts while still executing the
inner tool. The tool also runs in shadow mode when the decision request fails (a timeout, an HTTP
error, no connection): the guard logs a warning and passes the failure to on_decision as error,
with decision=None. Enforcement is fail closed: evaluation errors do not silently fall through to
the tool.
Compatibility: Python 3.10+, AutoGen Core 0.4–0.7; legacy AutoGen AgentChat 0.2 on Python 3.10–3.12.
Metadata
Release files for decionis-autogen 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| decionis_autogen-0.2.0.tar.gz | 8.7 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| decionis_autogen-0.2.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 16.6 kB
Release files / decionis_autogen-0.2.0.tar.gz
| Download URL | decionis_autogen-0.2.0.tar.gz |
|---|---|
| Size | 8.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
2b17bee9c0b5141624e9abe565aad32d661f593c0ff179a8b65b59c9975331b2
|
|
BLAKE2b-256 checksum How to use checksums |
28a43b37f6f2e677985e4fea2bc25dd9727ceeed6ac8e3254ef0157f74e46455
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.10
|
Release files / decionis_autogen-0.2.0-py3-none-any.whl
| Download URL | decionis_autogen-0.2.0-py3-none-any.whl |
|---|---|
| Size | 7.9 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
427f907b51cfa89f71da4c9dd25f783ca76467e9eb42caa43ba97ea6cdfa31de
|
|
BLAKE2b-256 checksum How to use checksums |
9ba02fa2c26434fb1cf484a3e86a9fac064ac58a8b2145c27369776d72f73915
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.10
|
Security for decionis-autogen 0.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.