# AI use cases for Jev Review (NiazMorshed2007/jev-review) Repo: [NiazMorshed2007/jev-review](https://github.com/NiazMorshed2007/jev-review) | Use cases: 3 | Page: https://dedupedai.com/use-cases/niazmorshed2007-jev-review/ Jev Review is a local MCP server that gives AI coding agents structured, per-dimension quality scores for their code changes, powered by the Jev API. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Score Claude Code changes for quality during a task: https://dedupedai.com/use-cases/niazmorshed2007-jev-review/score-claude-code-changes-with-jev/ - Score Codex changes against quality dimensions: https://dedupedai.com/use-cases/niazmorshed2007-jev-review/score-codex-changes-with-jev/ - Refactor in scored steps until deltas plateau: https://dedupedai.com/use-cases/niazmorshed2007-jev-review/refactor-until-dimension-scores-stop-improving/ ## How to score Claude Code changes for quality during a task with Jev Review Repo: [NiazMorshed2007/jev-review](https://github.com/NiazMorshed2007/jev-review) | Difficulty: Intermediate | Time: 15 minutes | Basis: From the README Install Jev Review in Claude Code so the agent scores its own diff at each checkpoint. **Who it is for:** A developer using Claude Code on a nontrivial feature who wants quality feedback before handoff. **The problem:** Agents finish a change without a structured check on readability, complexity, security, or test quality. **The outcome:** Claude Code calls jev_review on its focused diff and reports per-dimension scores it can act on. ### Steps 1. Run `export JEV_API_KEY="your-key"` in the terminal you use to start Claude Code. 2. Run `npx plugins add NiazMorshed2007/jev-review --target claude-code`. 3. Restart Claude Code and run `/mcp` to confirm that `jev-review` is connected. 4. Ask Claude Code to use `jev-review` while implementing a nontrivial change. ### Prompt to paste into your agent ```text Using the NiazMorshed2007/jev-review MCP tool jev_review, score my change to by sending the task and focused diff. Report each applicable dimension score and confidence, then list the weakest dimensions. Verify that the tool call succeeded and that no score is missing without an applicable flag. ``` ### Caveat Jev Review never reads your repository on its own, so the agent must send the task, diff, or files for each review. Risk, credentials: Your JEV_API_KEY is read by the local server and sent in the Authorization header to the Jev API, so keep it out of shared shell history and files. Risk, credentials: Review context you send to jev_review leaves your machine for TypeSafe's Jev API, so do not include secrets or unrelated proprietary code. 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/niazmorshed2007-jev-review/score-claude-code-changes-with-jev/ ## How to score Codex changes against quality dimensions with Jev Review Repo: [NiazMorshed2007/jev-review](https://github.com/NiazMorshed2007/jev-review) | Difficulty: Intermediate | Time: 15 minutes | Basis: From the README Install Jev Review for Codex so it scores a change before the final handoff. **Who it is for:** A developer using Codex who wants a quality check on a coherent code slice before merging. **The problem:** Codex output can pass tests while still scoring poorly on maintainability or changeability, and nothing flags it. **The outcome:** Codex runs jev_review on the slice and reports scores and dimension weaknesses before handoff. ### Steps 1. Run `export JEV_API_KEY="your-key"` before starting Codex. 2. Run `npx plugins add NiazMorshed2007/jev-review --target codex`. 3. Restart Codex and run `/mcp` to verify the connection. 4. If you prefer manual setup, add the `[mcp_servers.jev-review]` block with `command = "node"` and `env_vars = ["JEV_API_KEY"]` to `~/.codex/config.toml`. ### Prompt to paste into your agent ```text Use the jev-review MCP tool jev_review in the NiazMorshed2007/jev-review server to evaluate the current diff for . Return the 1 to 10 scores and confidence values for each applicable metric, and list the weakest dimensions by priority. Verify that each returned score has a confidence value between 0 and 1. ``` ### Caveat The README gives no Codex-specific rules for which files jev_review may send, so you choose the diff and files each call includes. Risk, credentials: The Codex config passes JEV_API_KEY through env_vars, so the key must be set in the environment Codex inherits and kept private. 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/niazmorshed2007-jev-review/score-codex-changes-with-jev/ ## How to refactor in scored steps until deltas plateau with Jev Review Repo: [NiazMorshed2007/jev-review](https://github.com/NiazMorshed2007/jev-review) | Difficulty: Advanced | Time: 30 minutes | Basis: From the README Use previousEvaluation so the agent sees which dimensions improved or regressed. **Who it is for:** A developer or agent refactoring a module who needs evidence that each change helps. **The problem:** Refactors often trade one quality dimension for another, and a single snapshot does not show it. **The outcome:** The agent gets a baseline, makes small edits, rescores, and stops when changes no longer produce real gains. ### Steps 1. Run `npx plugins add NiazMorshed2007/jev-review --target cursor` or the matching command for your client, then restart it. 2. Call jev_review with the task and focused diff to establish a baseline. 3. Make the smallest justified change, run your own tests, and call jev_review again with previousEvaluation set to the baseline. 4. Read the per-metric deltas, improvements, regressions, and unresolved weaknesses, then repeat or stop. ### Prompt to paste into your agent ```text Refactor with the jev_review tool from NiazMorshed2007/jev-review. Call it for a baseline, then after each change call it again with previousEvaluation set to the prior result. Report the per-dimension deltas and any regressions, and confirm that the project's existing tests still pass after each step. ``` ### Caveat The README says the Jev API enforces its own token ceiling, so large refactors may need to be split into slices before scoring. Risk, credentials: Every rescoring call sends the submitted context to the Jev API, so repeated loops repeat that disclosure and use the same API key. 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/niazmorshed2007-jev-review/refactor-until-dimension-scores-stop-improving/