Versions:
SASS (Saki Agent Secure Stream) is an agent-native SSH security suite developed by Saki Studio, designed to provide an application-layer overlay protocol for authenticated remote command execution between AI agents. Positioned within the network security and secure remote access category, the software addresses use cases where autonomous agents need to issue commands, exchange output, and coordinate tasks over an authenticated and defensible channel, such as multi-agent orchestration, automated infrastructure management, and agent-to-agent collaboration across untrusted networks. The current release is version 1.4.7, representing the seventh release in the SASS protocol v1.4 line, with two versions available in total, and the package ships with both daemon and client components to support server-side and client-side deployment. The suite implements several distinctive mechanisms: Total Response Mapping ensures complete capture of command output, Safety Gradient enables progressive permission escalation so that agents gain privileges incrementally rather than all at once, and Transparent Branching supports parallel execution paths for concurrent operations. Security is further reinforced by seven RFC-standardized defense plugins, including a ChaCha20-Poly1305 cognitive challenge for authenticated interaction, a zero-allocation tarpit to slow hostile connection attempts, an ED25519 audit chain for tamper-evident logging, and a Vi Swap terminal decoy, among others. SASS also serves as the reference implementation of IETF Internet-Draft draft-sakistudio-sass-07, tying the software directly to a documented, standardized protocol specification. This combination of authenticated transport, layered defensive tooling, and standards alignment makes the suite suitable for developers and operators building agent-based systems that require verifiable, auditable, and hardened remote execution without relying on conventional human-oriented SSH workflows.
Tags: