fipsagents 0.4.0

Production-ready AI agent framework for FIPS/OpenShift environments

fipsagents

Production-ready AI agent framework for FIPS/OpenShift environments. Provides BaseAgent — a pure Python, async-first base class that handles LLM communication, tool dispatch, MCP connections, prompt loading, skill management, configuration, and lifecycle so your agent subclass stays small.

Install

pip install fipsagents

With optional MemoryHub support:

pip install fipsagents[memory]

Quick start

from fipsagents.baseagent import BaseAgent, StepResult

class MyAgent(BaseAgent):
    async def step(self) -> StepResult:
        response = await self.call_model()
        return StepResult.done(response.content)

import asyncio
asyncio.run(MyAgent().start())

What's included

• LLM client via litellm — one interface for 100+ providers (vLLM, LlamaStack, OpenAI, Anthropic, Azure, Bedrock)
• Two-plane tool system — @tool decorator with agent_only, llm_only, or both visibility
• MCP client via FastMCP v3 — connect to remote tool servers
• Prompt loading — Markdown with YAML frontmatter
• Skills — agentskills.io progressive disclosure
• Configuration — YAML with ${VAR:-default} env var substitution
• MemoryHub — optional persistent memory (dual-path: MCP for LLM, SDK for agent code)
• Protective patterns — max iterations, exponential backoff, rate limiting

Used by

This package is the shared framework for templates scaffolded by the fips-agents CLI:

• agent-loop — single-agent loop (step() in a loop)
• workflow — directed graph of nodes with typed state

License

Apache 2.0