dsh-auto-guard
Run the following command in DeepSeek Harness:
dsh plugin install Ayle5678/dsh-auto-guard
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Ayle5678/dsh-auto-guard in your DeepSeek Harness terminal to install this command guard plugin; the source repository is available at https://github.com/Ayle5678/dsh-auto-guard .
About this plugin
Working across workspaces with danger-full-access means one stray rm -rf or curl | bash can wipe out hours of progress, and existing LLM review pipelines burn tokens by sending full context for every single command.
dsh-auto-guard layers a lightweight safety net on top of full access: a tiered decision pipeline (static allow-lists, an absolute deny-list, compound-command splitting, directory-deletion re-confirmation) handles the majority of commands instantly, and only genuinely uncertain ones are escalated to an LLM judge. Multi-level caching, a 60-day history layer, and auto-generated learned rules keep review costs to roughly 1–4% of total spend.
Built for developers who heavily rely on full-access mode, operate across multiple workspaces, and want a practical guardrail that reduces anxiety and accidental damage without choking productivity or inflating token bills. It is a safety net, not a sandbox.
Use Cases
- Reviewing shell commands quickly and reliably while working across multiple workspaces in full-access mode
- Splitting compound and pipeline commands to isolate dangerous sub-commands instead of pass-or-block the whole line
- Cutting LLM review costs to roughly 1–4% of total spend via allow-lists, multi-level caches, and learned rules
Best For
- DSH developers who rely heavily on danger-full-access mode across multiple workspaces
- Individual users who want to reduce accidental damage without giving up full-access convenience
- Teams that are sensitive to token spend and want to optimize per-command review costs
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.