# AI use cases for DSH Jev Plugin (luobosibing2/dsh-jev-plugin)

Repo: [luobosibing2/dsh-jev-plugin](https://github.com/luobosibing2/dsh-jev-plugin) | Use cases: 4 | Page: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/

A community plugin for the native DeepSeek Harness (DSH) that uses Jev or Luna Decisions as a judgment model for skill selection, task checks, log trimming, approvals and stage navigation.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Check a DSH turn against its Todo before it stops: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/check-todo-turns-before-they-stop/
- Narrow a DSH skill catalog to the skills a task needs: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/rank-skills-before-the-agent-loads-them/
- Shorten long command logs and passing test output for the model: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/trim-long-logs-and-test-output/
- Review the stages of a recorded DSH session by turn: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/review-stage-navigation-of-a-session/

## How to check a DSH turn against its Todo before it stops with DSH Jev Plugin

Repo: [luobosibing2/dsh-jev-plugin](https://github.com/luobosibing2/dsh-jev-plugin) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README

Runs a completion check on ordinary root turns that used Todo, with up to 3 automatic supplements by default.

**Who it is for:** A developer using DSH on multi-step coding tasks who wants a second check that the final answer covers the original request and the Todo list.

**The problem:** An agent can stop with a Todo item unfinished or a requirement missed, and the final answer may not show it.

**The outcome:** Each checked turn gets a complete, omission, contradiction, needs-user or unknown assessment, and gaps are supplemented in the same turn.

### Steps

1. Follow the Install through the Web UI section to add the plugin to a jev profile, then run `dsh --profile jev`.
2. Open Jev, then Settings and features, and select Jev or Luna Decisions under Judgment model.
3. Check the full request address and credential reference for the saved connection.
4. Enable Completion checks and keep Maximum automatic supplement rounds at the default of 3 unless you need a different limit.
5. Run a DSH turn that uses Todo, then read the assessment and any supplement the check added.

### Prompt to paste into your agent

```text
Using the repo luobosibing2/dsh-jev-plugin, enable the Completion checks feature in a DSH profile named jev, then run one root turn that uses Todo against <target task>. Report the assessment value and whether any supplement rounds ran. Verify that a turn without Todo was skipped.
```

### Caveat

A complete result only means the report appears to meet the requirements; the check does not verify code, artifacts or execution facts.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/check-todo-turns-before-they-stop/

## How to narrow a DSH skill catalog to the skills a task needs with DSH Jev Plugin

Repo: [luobosibing2/dsh-jev-plugin](https://github.com/luobosibing2/dsh-jev-plugin) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README

Recommends a configured number of relevant skills and ranks file glob results, without changing what the main agent loads.

**Who it is for:** A developer whose DSH setup has a large skill catalog and who wants fewer, more relevant skill suggestions for each task.

**The problem:** A long skill list makes it harder for the main agent to pick the right skill for a specific task.

**The outcome:** Each task gets up to the configured number of skill recommendations above the minimum relevance probability, and glob results are reordered.

### Steps

1. Follow the Install through the Web UI section to add the plugin to a jev profile, then run `dsh --profile jev`.
2. Open Jev, then Settings and features, and select the judgment model and connection you want to use.
3. Enable Skill selection and File ranking, and set the recommendation count and minimum relevance in the feature settings.
4. Start a task that matches one of your installed skills and read the recommended skills shown for it.

### Prompt to paste into your agent

```text
Using the repo luobosibing2/dsh-jev-plugin, enable Skill selection and File ranking in a DSH profile named jev, then send the task <target task description>. Report the recommended skills and their order. Verify that the main agent still loads the original skill.
```

### Caveat

The judgment model's recommendations are not a correctness guarantee, and a successful judgment can recommend no skills at all.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/rank-skills-before-the-agent-loads-them/

## How to shorten long command logs and passing test output for the model with DSH Jev Plugin

Repo: [luobosibing2/dsh-jev-plugin](https://github.com/luobosibing2/dsh-jev-plugin) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README

Removes clearly unneeded progress lines and repeated notices from command output while keeping failures and summaries.

**Who it is for:** A developer whose DSH agent runs long builds or test suites and fills its context with routine output.

**The problem:** Long command output, mostly passing tests and repeated progress notices, uses up context that the agent needs for the actual failure.

**The outcome:** Clearly unneeded log lines are removed after a command returns, failures and summaries stay, and an original-output recovery reference remains.

### Steps

1. Follow the Install through the Web UI section to add the plugin to a jev profile, then run `dsh --profile jev`.
2. Open Jev, then Settings and features, and select the judgment connection.
3. Enable Long-log admission and Test-log admission.
4. Run a build or test command that produces long output and check that failures, summaries and named or slow tests are still visible.

### Prompt to paste into your agent

```text
Using the repo luobosibing2/dsh-jev-plugin, enable Long-log admission and Test-log admission in a DSH profile named jev, then run <target test command>. Report which passing details were removed and confirm the failures and summary remain. Verify that the original-output recovery reference is present.
```

### Caveat

Judgments about which log lines are unneeded are model decisions and are not guaranteed to match what a developer would keep.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/trim-long-logs-and-test-output/

## How to review the stages of a recorded DSH session by turn with DSH Jev Plugin

Repo: [luobosibing2/dsh-jev-plugin](https://github.com/luobosibing2/dsh-jev-plugin) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README

Classifies complete recorded model steps on request and links stages to their original trajectory evidence.

**Who it is for:** A developer reviewing a past DSH session who wants to see which stage each step belonged to, such as exploration or editing.

**The problem:** A long recorded session is hard to scan step by step, and it is unclear where each phase of the work began and ended.

**The outcome:** Completed turns get stage labels with confidence values and links to the original steps, while the session records stay unchanged.

### Steps

1. Follow the Install through the Web UI section to add the plugin to a jev profile, then run `dsh --profile jev`.
2. Open Jev, then Settings and features, and turn on the Stage navigation switch.
3. Open the Stage navigation tab in the Web client, which follows Trajectory.
4. Select a completed turn, or request the session's unanalyzed completed turns, and review the classification input and answer shown on the page.

### Prompt to paste into your agent

```text
Using the repo luobosibing2/dsh-jev-plugin, turn on Stage navigation in a DSH profile named jev, then classify the completed turns of the session <target session id>. Report the stage labels per turn and confirm the original session records are unchanged. Verify that tool arguments shown in the classification input are redacted.
```

### Caveat

Stage labels and confidence values do not establish tool success or classification accuracy.

Risk, credentials: Tool arguments are sent for classification only after credential redaction, and incomplete argument JSON is replaced with a marker, so check the input shown on the page before relying on it.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/luobosibing2-dsh-jev-plugin/review-stage-navigation-of-a-session/