dsh-sandbox-escalation-fix
Run the following command in DeepSeek Harness:
dsh plugin install inmny/dsh-sandbox-escalation-fix
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, run the command `dsh plugin install inmny/dsh-sandbox-escalation-fix`, with the source code available at https://github.com/inmny/dsh-sandbox-escalation-fix。
About this plugin
When using DeepSeek Harness, models occasionally generate invalid permission escalation requests, particularly when they already hold sufficient access. This leads to frequent "not strictly wider" errors that disrupt tool invocation workflows, wasting resources and potentially triggering unnecessary security approvals.
The dsh-sandbox-escalation-fix plugin effectively solves this by automatically ignoring invalid escalation requests that do not exceed the current session permissions. It intelligently identifies and cleans up outdated or redundant sandbox_permissions parameters, handles edge cases like missing justifications, while strictly preserving DSH's original approval and validation mechanisms to ensure security processes remain intact.
This plugin is ideal for administrators and development teams managing stable, secure DSH environments. It streamlines permission management, reduces error-induced interruptions, and allows models to execute tasks more smoothly within authorized bounds, thereby enhancing overall system reliability and user experience.
Screenshots
Use Cases
- When models repeatedly trigger not strictly wider errors with existing permissions.
- When redundant or outdated escalation parameters need cleanup in tool calls.
- To automatically handle invalid permission requests in security audit environments and avoid disruptions.
Best For
- Administrators needing stable DeepSeek Harness operations.
- Operations teams responsible for permission control and security compliance.
- Developers aiming to reduce errors and optimize tool invocation efficiency.
Related Plugins
Adds an Auto approval mode on the official workspace-write sandbox, classifying semantic risks, asking when ambiguous, and denying destructive operations.
Undo/rollback system for DSH: config & plugin snapshots, one-click undo/redo/restore, message-level revert, secret masking, safe mode, and offline WebUI/GUI/CLI rescue tools for when DSH won't start.
Packages all 87 SKILL.md files from upstream reverse-skill as a DeepSeek Harness plugin that auto-registers them for authorized reverse engineering, penetration testing, and security research.
Dockyard DSH is a native DeepSeek Harness plugin that unifies official OAuth/client sessions for Codex, Antigravity, Grok, Claude, and Cursor, providing account pool, model catalog, and quota status.