Metadata-Version: 2.5
Name: intent-envelope
Version: 0.0.0
Summary: Declared intent envelopes for AI agents: a schema for what an agent may do, enforced outside the model.
Project-URL: Homepage, https://securedai.org/agent-intent
Project-URL: Documentation, https://securedai.org/architecture
Project-URL: Repository, https://github.com/securedai/intent-envelope
Author-email: SecuredAI <contact@securedai.org>
License-Expression: Apache-2.0
Keywords: agents,ai,guardrails,intent,mcp,security
Classifier: Development Status :: 1 - Planning
Classifier: Intended Audience :: Developers
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Security
Requires-Python: >=3.9
Description-Content-Type: text/markdown

# intent-envelope

A declared intent envelope states what an AI agent is for: the tools it may call, the
data it may touch, who it may contact, what it may spend, and which actions always
need a named human. The envelope is written before the agent runs and checked in code
at every tool call, so a boundary does not depend on the model behaving.

This release is a placeholder that reserves the name while the specification is
written. It contains no working code yet.

- The approach: https://securedai.org/agent-intent
- The six-layer architecture it belongs to: https://securedai.org/architecture
- Source, once published: https://github.com/securedai/intent-envelope

Maintained by SecuredAI.org.
