⬢github Rust · 1.3K ★ +24 since we first saw it · pushed 6 h ago · MIT
archestra-ai/OpenAPPA
Deterministic guardrails that don't break agents
OpenAPPA is a deterministic security layer that sits between an AI agent and its tools. It tracks the sensitivity of data an agent reads, and before each tool call checks a declarative TOML policy to decide whether that data is allowed to reach that destination. Unlike probabilistic PII detectors, it produces the same decision for the same input, and it runs in-process or as a sidecar.
Why now: It's trending after a Hacker News post, largely due to its benchmark claims: zero successful attacks across 1,320 evaluations while completing 88–90% of tasks, compared unfavorably to Microsoft FIDES and Claude Code auto mode, plus a NeurIPS 2026 workshop paper and Claude Code integration.
Who it is for: Developers and security teams building AI agents who need to prevent sensitive data from leaking to unauthorized tools without hurting task completion.
guardrailsguardrails-aisecurity
Stars over our 17 snapshots: 1.2K to 1.3K, since 4 h ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/archestra-ai/OpenAPPA