jev-codes
CLI and agent commands auditing a git diff against YAML standards packs, scoring each hunk with Jev so agents fix only flags.
Overview
jev-codes audits a git diff against editable standards packs, scoring it hunk by hunk with Jev. Jev never writes code; it answers typed yes/no, score and choice questions with probabilities, and the coding agent fixes only the flagged hunks. The README reports that a 500-line diff audits in seconds for a fraction of a cent, with one parallel Jev call per hunk.
It runs via npx, onboarding inline by asking for and validating a TypeSafe key, saving it with restricted permissions, and offering to install plugins for supported coding-agent harnesses including Claude Code and Cursor. audit checks the staged diff by default, or against a base branch with a chosen pack, and with --json writes only JSON to stdout and exits 1 when findings reach a --fail-on level (high by default in CI). Every harness adapter runs the same audit loop and fixes only high or medium findings at the reported lines. Community packs can be listed, added and updated with a preview of question and threshold changes, never silently.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does this hunk violate this coding standard? | yes · no |
| Score | How severe is the violation?(inferred) | none · low · medium · high |
How it works
- Select a standards pack
- Split the diff into hunks
- One parallel Jev call per hunk
- Report findings; exit non-zero at fail-on level
- Agent fixes only flagged hunks
Strengths
- Cheap, fast audits of large diffs
- Editable, shareable YAML packs
- CI-friendly JSON and exit codes
Caveats
- Model judgments may miss or over-flag
- Diff content is sent to TypeSafe
Getting started
`npx @kushwho/jev-codes` to onboard, then `npx @kushwho/jev-codes audit`.At a glance
- Type
- Agent skill
- Status
- beta
- Language
- TypeScript
- Runtime
- Node CLI, agent commands/skills
- Has code
- yes
Use cases and integrations
How curated lists describe it
Audit your git diff against YAML coding-standards packs using TypeSafe's Jev model, from a CLI or your AI agent's command/skill.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Whether each diff hunk violates a coding standard, and how severely.
Does Jev fix the code?
No, the agent fixes only flagged hunks.
Can I write my own standards?
Yes, standards packs are editable YAML.
Related projects
Semantic code linter that reads each method with its callers and callees and asks decision models about defects, kind, line and severity.
CLIbetaTypeScriptLocal MCP server giving Claude Code, Codex, Cursor and OpenCode per-dimension Jev quality scores on their code while they work.
ServicebetaTypeScriptCoding-agent hook that asks Jev one yes/no per project rule on every edit or turn and makes the agent fix any violation.
HookbetaTypeScriptAdvisor for coding agents: git and Stop hooks flag test tampering, secrets and skipped hooks, with optional Jev review findings.
HookbetaLocal MCP review engine where each rule becomes a Jev YES/NO/N/A Choice over agent-supplied diff chunks, with located evidence.
MCP serverbetaTypeScriptFuzzy linter hook for Claude Code and Codex where Jev checks each written file against team rules and reports violations at edit time.
HookbetaTypeScript