dsh-aggression: Recommendations for Correcting Aggressive Behavior
The DeepSeek Harness (DSH) plugin dsh-aggression is an admin-security tool designed to help developers detect and correct "aggressive" or specific behavior patterns based on input parameters. It registers an `aggression` tool implemented locally in pure Node.js, enabling behavior validation without network access. Once installed, it can be invoked during interactive sessions, making it suitable for scenarios requiring local behavior intervention. However, users should review its permission scope before use.
Read Moredsh-toilet-train: Potty Training
The DSH plugin `dsh-toilet-train` focuses on providing behavioral correction advice. Maintained by uckkk and licensed under the MIT License, it is implemented in pure Node.js and does not rely on network requests. Installation requires specific environment dependencies. The plugin provides the `toilet_train` tool, which can return behavioral logic judgments based on parameters. Because it runs with DSH process permissions, users should review its source code and license before installation. It is a lightweight tool suitable for local use cases.
Read Moredsh-fear-pet: A Fear-Mitigation Tool That Returns Behavioral Correction Suggestions Based on Parameters
dsh-fear-pet is a DSH plugin maintained by uckkk, designed to provide lightweight behavioral correction suggestions. It runs entirely locally with no network dependencies, and its core function is to return specific correction recommendations based on input parameters. Once installed, its registered tool can be invoked during a session, making it suitable for scenarios that require behavioral review or correction of specific inputs. Before use, ensure the environment meets the relevant dependency requirements.
Read Moredsh-jumping-fix: Player Pounce Correction
The DeepSeek Harness (DSH) plugin dsh-jumping-fix is designed to address agent behavior consistency issues by providing runtime behavior correction capabilities. Maintained by uckkk, its core function is to optimize agent execution logic by returning correction suggestions based on parameters. It is implemented in pure Node.js, has no network dependencies, and has specific version requirements. Once installed, its tool can be invoked within a session. Since it runs with the permissions of the current process, users are advised to review the source code before installation.
Read More