Jev Review — Development skill for Claude Code
Review code and let Jev filter the findings before you report them.
How to install Jev Review
Installs to ~/.claude/skills/buchmark-claude-jev-jev-review/SKILL.md
mkdir -p ~/.claude/skills/buchmark-claude-jev-jev-review && curl -fsSL https://raw.githubusercontent.com/buchmark/claude-jev/HEAD/commands/jev-review.md -o ~/.claude/skills/buchmark-claude-jev-jev-review/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Jev Review does
description: Review code and let Jev filter the findings before you report them argument-hint: [diff | branch | file | PR number]
Review this target, defaulting to the uncommitted changes when none is given: $ARGUMENTS
- Read the change and collect candidate findings yourself. For each one write a short id, the file, the line, a one-sentence claim naming the defect, and the concrete inputs that produce the wrong result.
- Call
jev_review_findingswith those findings, the `sou
Alternatives in Development
- /validate — Run full validation on the current finding before writing a report 927 ★
- Synthesize Findings — Combine findings from multiple research agents into a coherent, well-structured research report 228 ★
- Governance Review Agent — Chains adoption-report, drift detection scoped to at-risk teams, and stakeholder-brief into a periodic governa 194 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Rocket Fuel Skill
Run Claude Fable 5 + OpenAI Codex as co-founders in one Claude Code skill. Fable is the Visionary (plan, stand
Validate Agent
Gate a single agent finding through the FP gate + Negative-Result gate before writing any report. Kills findin
Claim Review
Run the three reviewer agents in parallel over a scope and consolidate their findings into one report.
Jev Skill
Jev skill for Claude Code and Codex: TypeSafe System One guide, Python CLI, 5 tested recipes, and findings fro
Econ Paper Review Skill
Referee report on your economics paper before a real referee sees it — verified comments and a revision plan y
Fable Eval
Run the fable-mode eval suite (probes → pairwise judge → report). Costs tokens — runs headless claude many tim
Related Agents
Filter Tuner
Looks at what the deterministic filter and the AI scoring kept or rejected, finds the rules that let bad offer
Review Bug
Review: Bug/Logic Error Detection (coverage-first) — report all findings with confidence and severity tags; do
Squall
Use for an adversarial bug, crash and defect review of a diff before a PR is ready, or to triage a bug report.