kilo-zen2dsh
Run the following command in DeepSeek Harness:
dsh plugin install Xyanxhu/kilo-zen2dsh
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Xyanxhu/kilo-zen2dsh in the DeepSeek Harness terminal to install this plugin; source code is available at https://github.com/Xyanxhu/kilo-zen2dsh
About this plugin
Running DSH agent turns against a local LLM has traditionally meant either paying for a paid API endpoint or standing up a sidecar process that binds a local port. kilo-zen2dsh sidesteps both: it registers two native LlmAdapters—the Kilo Gateway and the OpenCode Zen free lane—directly into the DSH route table as keyless channels. No child process, no local bind, no default credential; the model picker simply shows a curated set of free options the moment DSH starts.\n\nThree capabilities carry the core workload. Dynamic free-model discovery reads the isFree or is_free flag when present and falls back to :free suffix matching, backed by a seven-day disk cache and a periodic refresh so the catalog stays current. A deliberately conservative model filter keeps only text-output models that advertise tool support, which is exactly what a DSH agent turn requires. Finally, automatic context and output-limit reconciliation merges inconsistent upstream limit declarations and applies a per-model ceiling immediately before the wire request, catching cases like MiniMax-M3 where a 1 M-token window is silently rejected above 524 K.\n\nThe plugin suits solo developers and individual researchers who want a zero-cost chat or lightweight agent loop inside DSH without provisioning any infrastructure. Both lanes are best-effort promotional tiers: upstream quota, rate limits, and model availability remain under provider control, and nothing in the plugin bypasses authentication or billing on either gateway.
Use Cases
- Call Kilo or Zen free models directly from the DSH model picker without configuring any API key
- Run lightweight agent turns at zero cost using free models with text output and tool support
- Skip sidecar processes and local port binding by connecting to upstream gateways as a pure DSH adapter
Best For
- Solo developers who want to try DSH agent loops at zero cost
- Researchers who prefer to avoid deploying an extra sidecar process for local inference
- Individual users who want to test model behavior within free tiers without a paid account
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
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.