dsh-auto: Adding Auto Approve Permission Presets for Model Review in DSH Web UI
dsh-auto is an MIT-licensed plugin for the DSH Web UI, designed to optimize the approval process. It introduces a constrained sub-Reviewer Agent to perform automated review before human confirmation, reducing manual intervention for repetitive or low-risk operations. The plugin supports only the Web UI. Its core mechanism is that when a session selects `Auto Approve`, a Reviewer running in a read-only sandbox is launched. The Reviewer is restricted to a whitelist of tools (such as read, grep), is prohibited from spawning further sub-agents, and has a step limit, enabling bounded read-only investigation and producing structured results. The host process makes the final decision based on these results: it rejects all critical risks, also rejects high risks that do not meet authorization conditions, and defaults to shutdown in case of failure. Installation requires adding the plugin via a command and restarting, with optional custom configuration.
Read Moredsh-auto-approve: Adding Auto Permission Level Between Workspace Write and Full Access
dsh-auto-approve inserts an auto permission tier between workspace-write and danger-full-access for DeepSeek Harness. The sandbox boundary is identical to workspace-write, with routine sandbox escalations approved in one go by a classification model; cases hitting a deterministic danger list, where the classifier is uncertain, or due to timeout or malformed responses are still escalated to manual approval. The plugin supports Web, TUI, and Desktop frontends and can be installed via dsh plugin and enabled by activating the Auto tier in Permissions.
Read More