
hellosky983/dsh-skillradar
30Last commit Aug 14, 2026
dsh-skillradar DSH plugin
dsh-skillradar analyzes every skill visible to the current DSH session, scores each against the recent conversation text (English + Chinese token overlap), and returns a ranked recommendation of which skill to load next. It provides a fast, deterministic relevance signal for agents and users working across many skill domains. The plugin is dependency-free, local-only, and requires no network or credentials.
How to install the dsh-skillradar DSH plugin
dsh plugin add github:hellosky983/dsh-skillradarCopying does not run this command. Review the repository and version before installing the dsh-skillradar DSH plugin.
dsh-skillradar DSH plugin data source
dsh-skillradar DSH plugin snapshot date: Aug 16, 2026
discovered
What the dsh-skillradar DSH plugin can do
- Scans all skills visible to the current session
- Scores each skill against recent conversation text using English and Chinese token overlap
- Returns a ranked list of recommended skills with percentage scores
- Configurable session_id and limit parameters
- All scoring is local, no network calls or external services
Where the dsh-skillradar DSH plugin fits
- When the user or agent is unsure which skill fits the current task
- For workflows that span multiple skill domains and need a quick relevance signal
- To automatically trigger skill loading based on conversation context
- As a deterministic fallback when model-based selection is unreliable
Who the dsh-skillradar DSH plugin is for
- DSH users who manage many skills and want a tool to pick the best one
- Developers building agents that need to dynamically select skills
dsh-skillradar DSH plugin limitations
- Only supports English and Chinese token overlap scoring
- Scans only skills visible to the current session (not all available skills)
- Requires DSH version 0.1.0-rc.6 or later
- Plugin must declare `inject = ['tools']` (fixed in latest commit)
dsh-skillradar DSH plugin: from the repository README
Quoted from the hellosky983/dsh-skillradar README, the upstream source of the dsh-skillradar DSH plugin. Copyright remains with the original authors.
A DeepSeek Harness plugin that scans every skill visible to the current session, scores each against the recent conversation text (English + Chinese token overlap), and returns a ranked recommendation of which skill to load next. ## Overview DSH sessions expose many skills, and neither the model nor the user always knows which one applies. This plugin answers "what should I load right now?" with a concrete, ranked answer. It suits agents and users who work across many skill domains and want a fast, deterministic relevance signal. ## Compatibility - DSH version: `0.1.0-rc.6` - Mainline: verified against `deepseek-harness` mainline snapshots of 2026-08-14 - Last verified: 2026-08-14 ## Install / Uninstall Install (from GitHub): ```bash dsh plugin add github:hellosky983/dsh-skillradar ``` Or clone and install locally: ```bash git clone https://github.com/hellosky983/dsh-skillradar.git cd dsh-skillradar dsh plugin add . ``` Upgrade: re-run the install command after `git pull`. Disable temporarily: remove the plugin row from your profile composition, or run: ```bash dsh plugin remove dsh-skillradar ``` Uninstall: remove the `dsh-skillradar` dependency and its bundle entry f
Read the full READMERepository license: MIT
dsh-skillradar DSH plugin questions
How do I install dsh-skillradar?
Run `dsh plugin add github:hellosky983/dsh-skillradar` in your terminal. Alternatively, clone the repo and run `dsh plugin add .` from the cloned directory. After installation, restart DSH to activate the plugin.
What does the Skill Radar tool actually do?
It scans all skills visible to the current session, scores each one against the recent conversation text (using English and Chinese token overlap), and returns a ranked list with percentage scores. The tool is invoked via `skill_radar` with no arguments, or you can specify a session_id and limit.
Does the plugin send data to any external server?
No. All scoring is performed locally using text analysis. The plugin reads session messages and skill metadata, but never makes network requests. No credentials or environment variables are required.
The tool is not showing up after install. What should I do?
Restart DSH first. If it still doesn't appear, run `dsh --dump-config` to check if the plugin bundle row is listed. Also ensure that your DSH version is at least 0.1.0-rc.6. If you see an error about 'cannot get property tools without inject', reinstall the latest commit which properly declares `inject = ['tools']`.
Can I use this plugin with languages other than English and Chinese?
The scoring algorithm is based on token overlap for English and Chinese only. It may not work effectively for other languages because the token matching logic is designed for those two languages. The plugin does not currently support multilingual scoring.