sentinel-agent-trust
Python client SDK for Sentinel AI — the identity & governance layer for AI agents: control, authorize, and audit autonomous agents before they touch your production APIs.
Sentinel is in private beta. For source access, a hosted sandbox organization key, or to collaborate: cosimo@plannest.net
Install
pip install sentinel-agent-trust
Usage
import os
from agent_trust_sdk import AgentTrustClient, AgentPermissionDeniedException
client = AgentTrustClient(
agent_id="agt_...",
private_key=os.environ["SENTINEL_AGENT_PRIVATE_KEY"], # base64 Ed25519 seed
gateway_url="https://your-gateway.example.com",
)
try:
client.enforce_action("crm.refund.create", "crm://refunds", {"amount": 90})
crm.create_refund(amount=90) # authorized → do the real thing
except AgentPermissionDeniedException as exc:
# final denial: do NOT retry — branch (degrade, escalate to a human, give up)
escalate_to_human(exc.reason, exc.request_id)
Every request is signed with the agent's private Ed25519 key (never sent to
the server) and carries a timestamp + nonce, so requests cannot be replayed.
Denials carry a machine-readable code (permission_denied,
behavioral_limit_exceeded, rate_limited, …) so your agent can branch
instead of looping.
Proxy mode
With proxy mode the request travels through the gateway, which authorizes it, injects the upstream credential (the agent never holds it) and forwards it — enforcement the agent cannot skip:
resp = client.proxy_request("stripe", "POST", "v1/refunds",
payload={"charge": "ch_123", "amount": 90})
print(resp.status_code, resp.json()) # the upstream's response
Method, path, query and body are all covered by the Ed25519 signature.
Denials raise the same exceptions as enforce_action; a pending human
approval raises AgentApprovalPendingException — retry the identical call
after approval.
Generate a keypair
python -m agent_trust_sdk.keygen
Register the public key with your Sentinel gateway (POST /v1/agents); keep
the private seed in a secret store.
License
Apache 2.0 — see LICENSE and NOTICE.
Metadata
Release files for sentinel-agent-trust 0.3.0
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 | Actions |
|---|---|---|---|---|
| sentinel_agent_trust-0.3.0.tar.gz | 13.5 kB | With provenance | Details for sentinel_agent_trust-0.3.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| sentinel_agent_trust-0.3.0-py3-none-any.whl | Python 3 | none | any | Details for sentinel_agent_trust-0.3.0-py3-none-any.whl |
Total release size: 27.7 kB
Release files / sentinel_agent_trust-0.3.0.tar.gz
| Download URL | sentinel_agent_trust-0.3.0.tar.gz |
|---|---|
| Size | 13.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
8d651d6aa948ba6b73d8ecd8e0ad29595dbd52cebbe709029d58f3b6d1c9a982
|
|
BLAKE2b-256 checksum How to use checksums |
faa11bba054833313c9c6fd3b06585ef136cb5cb704aa8843c2e42efbdfe3dfe
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
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.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 16, 2026.
Transparency logRelease files / sentinel_agent_trust-0.3.0-py3-none-any.whl
| Download URL | sentinel_agent_trust-0.3.0-py3-none-any.whl |
|---|---|
| Size | 14.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
c8881a8f3c87f52751e3601ef32a8ec967f981c04e72c4a06bfe363b2942b09b
|
|
BLAKE2b-256 checksum How to use checksums |
658506da5d810feb2f5a41dd9652f59fbd492b9c5a96ccc8e48a9e1c8d4d4d61
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
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.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 16, 2026.
Transparency logSecurity for sentinel-agent-trust 0.3.0
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.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for sentinel_agent_trust-0.3.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 16, 2026.
Transparency logA 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.
To learn more about protecting yourself from malicious actors, visit our package security documentation.