dsh-trash
Run the following command in DeepSeek Harness:
dsh plugin install LeonSone/dsh-trash
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install LeonSone/dsh-trash in DeepSeek Harness to install; the source repository is at https://github.com/LeonSone/dsh-trash. Restart dsh web after installation to load the plugin.
About this plugin
When you grant your AI agent danger-full-access and let it bulk-delete files, a single rm -rf or Remove-Item can wipe days of work with zero chance of recovery. dsh-trash exists for exactly this scenario: it intercepts every delete command at the tool pipeline pre-execute gate, converting each deletion into a recoverable move-to-trash that you can undo with a single call.
The plugin ships four tools: trash (move files or directories into the trash), trash-list (inspect entries with timestamps and sizes), trash-restore (recover an entry to its original path), and trash-purge (the only path that truly deletes, and it demands an explicit confirm flag). The detector follows a fail-safe philosophy — it prefers false positives over letting a real delete slip through — while short aliases like rm or del only match in command position, so tutorial strings and grep patterns stay untouched.
It is built for anyone running pwsh, bash, or run_code tools inside DeepSeek Harness who needs file operations and wants a second chance on every deletion, especially at the highest permission tier. Zero runtime dependencies, drop-in installation, and no workflow changes required.
Use Cases
- Granting full-access to an AI agent for bulk file cleanup with zero tolerance for data loss
- Needing a safety net when the agent runs rm, Remove-Item, or shutil.rmtree
- Performing periodic permanent cleanup without the risk of irreversible mistakes
Best For
- Developers running file operations at danger-full-access in DeepSeek Harness
- Engineers delegating bash, pwsh, or run_code tasks to an AI agent
- Users building automated workflows that demand a file-deletion safety net
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.