Runtime safety for AI agents: intercept tool calls, policy scoring, and audit logging
Project description
Agentiva
Runtime safety for AI agents. Agentiva helps you:
- intercept tool calls (before side effects)
- score risk against a YAML policy
- run in shadow mode first, then live (block) or approval
- keep an audit trail of what happened
Install
pip install agentiva
Use as a repo scanner (optional git hook)
agentiva scan .
agentiva init # installs a pre-push hook that runs `agentiva scan .`
View the local HTML report generated by scans:
agentiva dashboard
Use at runtime (agent frameworks / custom)
Run the API:
agentiva serve
Wrap tools in your app (minimal example):
from agentiva import Agentiva
shield = Agentiva(mode="shadow")
tools = shield.protect([your_tool_a, your_tool_b])
Project links
- Source + full docs: https://github.com/RishavAr/agentiva
- Issues: https://github.com/RishavAr/agentiva/issues
Note about the dashboard UI
The PyPI distribution installs the Python package (CLI + API + policies). The full Next.js dashboard UI lives in the repository under dashboard/ and is not included in the wheel.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file agentiva-0.1.5.tar.gz.
File metadata
- Download URL: agentiva-0.1.5.tar.gz
- Upload date:
- Size: 148.9 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
06160485a7f96c52be9fa56b40621def461d619dbcefd0010a6363442546814f
|
|
| MD5 |
702f264dbc0d99cd33cfdd71bd35439f
|
|
| BLAKE2b-256 |
172405076765a0b5eba9043c68210166bd9e51e873cc5c1870c3e5e03e9cab0d
|
File details
Details for the file agentiva-0.1.5-py3-none-any.whl.
File metadata
- Download URL: agentiva-0.1.5-py3-none-any.whl
- Upload date:
- Size: 125.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
10d8377b2f085bef06a658f5a5082f6f6efdd222d94572d91ea1f2e4bc602212
|
|
| MD5 |
ac4ac8829c76a8a58bbf9375242182c4
|
|
| BLAKE2b-256 |
0423be2fbc6a7a30c25d53cb5e9447d3cbe5cc979116c06dde74b41323eee9d3
|