Skip to main content

SDK for building agents that run within Sage Sanctum infrastructure

Project description

Sage Sanctum Agent SDK

SDK for building agents that run within the Sage Sanctum secure multi-agent execution platform.

Features

  • SPIFFE Authentication - Automatic JWT SVID management for agent identity
  • Transaction Tokens (TraT) - IETF-standard authorization for scoped transactions
  • LLM Gateway Integration - Route LLM calls through authenticated, policy-enforced gateways
  • Multi-Provider Support - OpenAI, Anthropic, and Google via unified interface
  • SARIF Output - Standard static analysis output format for GitHub Code Scanning
  • Testing Utilities - Mock gateway, LLM, and TraT clients for unit testing

Quick Start

from sage_sanctum import AgentContext, AgentRunner, SageSanctumAgent, AgentResult
from sage_sanctum.io.inputs import AgentInput, RepositoryInput
from sage_sanctum.io.outputs import SarifOutput, Finding, Location
from sage_sanctum.llm.model_category import ModelCategory


class MySecurityAgent(SageSanctumAgent):
    @property
    def name(self) -> str:
        return "my-security-agent"

    @property
    def version(self) -> str:
        return "0.1.0"

    async def run(self, agent_input: AgentInput) -> AgentResult:
        # Get an LLM client for analysis
        llm = self.context.create_llm_client(ModelCategory.ANALYSIS)

        # Your agent logic here...
        response = llm.invoke([...])

        return AgentResult(
            output=SarifOutput(
                tool_name=self.name,
                tool_version=self.version,
                findings=[...],
            ),
            exit_code=0,
        )


# Entry point
if __name__ == "__main__":
    import sys
    sys.exit(AgentRunner(MySecurityAgent).run())

Installation

pip install sage-sanctum-sdk

Architecture

Agents run in isolated containers with no direct network access. All external communication (LLM calls, MCP tools, packages) flows through authenticated gateways:

Agent Pod (seccomp: no AF_INET)
  └── Agent SDK
       ├── SPIFFE JWT (identity)
       ├── Transaction Token (authorization)
       └── Gateway Client (Unix socket)
            └── LLM Gateway → OpenAI / Anthropic / Google

License

MIT

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

sage_sanctum_sdk-0.3.0.tar.gz (73.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

sage_sanctum_sdk-0.3.0-py3-none-any.whl (45.1 kB view details)

Uploaded Python 3

File details

Details for the file sage_sanctum_sdk-0.3.0.tar.gz.

File metadata

  • Download URL: sage_sanctum_sdk-0.3.0.tar.gz
  • Upload date:
  • Size: 73.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for sage_sanctum_sdk-0.3.0.tar.gz
Algorithm Hash digest
SHA256 fa573060419b3ca7fc10115a398b86f79fb83a4d638bdb262fea83fad4e469fe
MD5 0ea83656002d3d1ec0ecc7b52bdddc2e
BLAKE2b-256 2c91900ac18e07e4214780053ff1c0433efa231b159f399429fd2e753e26e64b

See more details on using hashes here.

File details

Details for the file sage_sanctum_sdk-0.3.0-py3-none-any.whl.

File metadata

File hashes

Hashes for sage_sanctum_sdk-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 547615f07f6dce67931f09c38a9373754734b5b3700bbb316e5559dfb27e602f
MD5 63efc358481a09dd3421686724d1c9e9
BLAKE2b-256 8fc2160d3fe77c3eeaf1b295a3822aaee2fd6c7c9f247510af4de97523f68a72

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page