dsh_plugin_file_attach
Run the following command in DeepSeek Harness:
dsh plugin install TestTheBoy/dsh_plugin_file_attach
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install TestTheBoy/dsh_plugin_file_attach in your terminal, or visit https://github.com/TestTheBoy/dsh_plugin_file_attach to obtain the source and install from the local directory. After installation, restart dsh web and refresh your browser to see the file-attach button.
About this plugin
When you use DSH Web for technical Q&A or code review, the most tedious step is rarely the question itself—it is hand-copying local config files, logs, or code snippets into the prompt box. Long content gets truncated, and every revision means pasting all over again. dsh_plugin_file_attach turns that into a single click: a capsule button appears in the input toolbar, and attached files show up as removable chips above the input, scrollable horizontally when they exceed one line, without interfering with each other.
The core capability is automatic context injection. UTF-8 text, code, Markdown, JSON, and log files are all supported, along with PNG, JPG, WebP, and GIF images. On every message you send, the host injects the attachment content into the model context during the pre-step phase—no manual pasting, no extra file-read tool calls. The limits are clear: 256 KB per file, 10 files at most, 1 MiB total. Binary or unreadable files are rejected with a reason. One important detail: attachment content is injected only for the current request and is never written into chat history, keeping your conversation record clean.
If your daily work involves having the model cross-reference local configs for debugging, reading through code snippets, interpreting service logs, or analyzing screenshots, this plugin removes the copy-paste friction and lets you focus on the actual problem.
Use Cases
- Send local JSON configs or log files directly to the model instead of manual pasting
- Let the model reference code snippets in-context without repeated copy-paste
- Upload screenshots for the model to identify and explain configs or errors
Best For
- Developers who rely on DSH Web for day-to-day technical Q&A
- Ops engineers who frequently feed local files or service logs to the model
- Engineers doing code review or config troubleshooting in the browser
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.