DeepSeek Harness plugin

dsh-llm-friendli

FriendliAI serverless LLM adapter for DeepSeek Harness — OpenAI-compatible chat completions with dynamic model discovery and reasoning support

Jump to install

Source facts

Repository
Lee-Si-Yoon/dsh-llm-friendli
Latest update
Aug 18, 2026
Category
Models & Providers
GitHub stars
0
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/Lee-Si-Yoon/dsh-llm-friendli
Plugin: dsh-llm-friendli
Author: Lee-Si-Yoon

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer3 files
README.mdSource · read only

dsh-llm-friendli

![npm](https://www.npmjs.com/package/dsh-llm-friendli)

A DeepSeek Harness plugin that adds FriendliAI as an LLM provider. Friendli speaks the OpenAI-compatible chat API, so the adapter implements the harness LlmAdapter contract over it and fetches the model catalog from Friendli at runtime.

Quick start

Install the dsh CLI, add the plugin to a profile, export your key, and launch:

npm i -g @deepseek-ai/dsh                       # the dsh CLI (needs Node.js)
dsh plugin --profile web add dsh-llm-friendli   # register the friendli provider
export FRIENDLI_API_KEY="flp_..."               # your Friendli key
dsh web                                         # friendli is now a provider

That's it. dsh plugin add merges the plugin's bundle patch into the profile, so the friendli provider is registered with working defaults — no config file to touch. Open the web UI, pick a Friendli model, and go. The model list is pulled live from Friendli.

Prefer not to install globally? Prefix each command with npx @deepseek-ai/dsh instead.

Confirm the layer loaded without booting:

dsh --profile web --dump-config   # look for a "# == dsh-llm-friendli" layer

Get a key from the Friendli dashboard. It's read from the environment on every request, so never inline it in a config file or commit it.

Configuration

The defaults work out of the box. To override any of them, add a config block in your profile's own cordis.patch.yml:

- id: dsh-llm-friendli
  name: 'dsh-llm-friendli'
  config:
    apiKeyEnv: FRIENDLI_API_KEY                     # default
    baseURL: https://api.friendli.ai/serverless/v1  # default
    providers: [friendli]                           # default
    thinking: enabled                               # optional; on/off for controllable models
    modelCacheTtlMs: 60000                          # default; model-catalog cache

To use it outside dsh web, point an agent at a Friendli model id (as returned by GET /models). See [examples/cordis.yml](examples/cordis.yml) for a full fragment.

- id: agent-loop
  name: '@deepseek-ai/dsh-agent-loop'
  config:
    agents:
      - id: main
        provider: friendli
        model: zai-org/GLM-5.2

License

MIT © Lee-Si-Yoon