Agentic Permissions Policy Algebra (OpenAPPA)
A declarative policy format and deterministic engine that labels data flowing between an LLM agent and its tools, so restricted data cannot reach audiences it was never allowed to reach.
Canonical text: https://www.openappa.com/contracts, read against version v0.37.0.
Core concepts
- An OpenAPPA policy defines restrictions on tool results and requirements for tool calls.
- An audience identifies who can receive data.
- Trust describes how much OpenAPPA can rely on the data the agent has received.
- Effects record successful actions.
- Attention requires fresh approval for each call.
- An annotator classifies a tool call.
- A sanitizer cleans or validates data before an agent or tool receives it.
- An authority reviews a tool call that would otherwise be blocked.
- Remedy plans give the agent ways to continue when a call is blocked or a result would add restrictions.
Facts
- Maintainer
- Archestra.AI
- Repository
- https://github.com/archestra-ai/OpenAPPA
- Licence of the text
- MIT
- First published
- 2026-08-18
- Status of the text
- draft
- Listing
- listed, not yet claimed
Read with
- claimed
Purpose-bound Access Control and Trust
How an AI agent proves who it acts for, with what declared purpose and evidence, to a gateway that already authenticates and authorizes; a PACT pass never grants access and a native denial always wins.
v1.0.0-alpha.0 Agentic Stack - claimed
Agent Operation Protocol
Six laws a system must enforce at its API boundary so that any actor, human or agent, can act on a shared operational graph without breaking it.
-
Model Context Protocol
An open protocol, built on JSON-RPC, for connecting LLM applications to external data sources and tools through hosts, clients and servers.
Badge
[](https://manifesto.md/openappa/) Maintainers claim a listing with a pull request setting listing: claimed, or a yes on the listing issue.