dsh-lens-lite
Run the following command in DeepSeek Harness:
dsh plugin install ben7am1n/dsh-lens-lite
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install ben7am1n/dsh-lens-lite in DeepSeek Harness to install the plugin (source: https://github.com/ben7am1n/dsh-lens-lite).
About this plugin
In DeepSeek Harness, dsh-lsp provides semantic navigation—goto definition, find references, hover—yet nobody tells the model "this file has a type error" right after an edit. The model writes code, then discovers the error only a few turns later, almost by accident. dsh-lens-lite closes that gap: after any successful file-mutating tool call, it runs the type-checkers or linters you have configured and attaches their output to the same tool result, so the very next model request already carries the diagnostics.
The plugin ships zero built-in commands. Every checker is yours to declare in config—which executable to run, which file extensions it claims, how long it may take, and how much output to keep. Arguments support placeholders like {file} and {dir} for precise, single-file invocations. All matching checkers run concurrently, and their combined output arrives as a compact, collapsible notice in the transcript rather than a wall of compiler text flooding the context window. The plugin never vetoes or rewrites a tool call; it simply folds diagnostic context onto whatever decision came back from the downstream waterfall.
If you use DeepSeek Harness for day-to-day code editing and want the model to catch mistakes on the very first turn after a write—instead of discovering them three turns later—dsh-lens-lite is a lightweight, zero-intrusion fit. It only requires that your toolchain has a command you can invoke per file (tsc, ruff, golangci-lint, or any equivalent). For large repositories, prefer single-file invocation patterns to keep per-edit latency low, and raise timeoutMs only for checkers that genuinely need it.
Use Cases
- Automatically run tsc or ruff after a file write so the model sees type errors on the very next turn
- Configure multiple checkers per file extension and let the plugin dispatch them concurrently on every edit
- Use single-file lint invocations on large repos to keep per-edit latency predictable
Best For
- Developers doing day-to-day code editing with DeepSeek Harness
- Tech leads who want to collapse the multi-turn error-discovery loop into immediate post-edit feedback
- Harness users who have a type-checker toolchain but lack a post-edit diagnostics hook
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.