← 返回 Skills 市场
anderskev

Review Remix V2

作者 Kevin Anderson · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
15
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install review-remix-v2
功能描述
Comprehensive Remix v2 code review with per-area review skills, run in parallel where the agent supports subagents and sequentially otherwise. Detects Remix...
使用说明 (SKILL.md)

Remix v2 Code Review

Arguments

  • --parallel: Hint to fan out per technology area when the agent supports subagents (see Step 5). When unsupported, the review runs sequentially with identical output.
  • Path: Target directory (default: current working directory)

Step 1: Identify Changed Files

git diff --name-only $(git merge-base HEAD main)..HEAD | grep -E '\.(tsx?|jsx?|mjs|cjs|css)$|^(remix|vite)\.config\.'

If nothing is returned, ask the user for an explicit path list before continuing.

Step 2: Detect Remix v2 + Techs

# Detect Remix v2 in package.json (any official adapter)
grep -E '"@remix-run/(react|node|cloudflare|deno|serve)"' package.json | head -3

# Confirm app router layout
ls app/routes/ 2>/dev/null | head -5

# meta exports (route metadata)
grep -rE 'export const meta|export function meta|export let meta' app/ --include="*.tsx" --include="*.ts" -l | head -3

# Sessions / cookies / auth
grep -rE 'createCookieSessionStorage|getSession\(|commitSession\(|destroySession\(' app/ --include="*.ts" --include="*.tsx" -l | head -3

# Form primitives (forms-review)
grep -rE 'useFetcher\(|\x3CForm|useSubmit\(|useNavigation\(' app/ --include="*.tsx" -l | head -3

# Form anti-patterns: native form / manual fetch (forms-review trigger even without Remix imports)
grep -rE '(^|[^.])\x3Cform |fetch\(' app/ --include="*.tsx" -l | head -3

# Error boundaries
grep -rE 'export function ErrorBoundary|export const ErrorBoundary|useRouteError|isRouteErrorResponse' app/ --include="*.tsx" -l | head -3

# Headers / streaming / server-only modules
grep -rE 'export const headers|export function headers|defer\(' app/ --include="*.tsx" --include="*.ts" -l | head -3
find app -type f \( -name '*.server.ts' -o -name '*.server.tsx' -o -name '*.client.ts' -o -name '*.client.tsx' \) | head -5

# Prefetch hygiene (perf-ssr-review)
grep -rE 'prefetch=|\x3CPrefetchPageLinks' app/ --include="*.tsx" -l | head -3

# Loaders / actions (data flow)
grep -rE 'export (async )?function (loader|action)|export const (loader|action)' app/ --include="*.tsx" --include="*.ts" -l | head -3

If the package.json check returns nothing, stop and tell the user: this skill expects a Remix v2 project.

Step 3: Load Verification Protocol

Load the review-verification-protocol skill before any substantive judgment on code. This is the canonical cross-plugin protocol; do not substitute a framework-specific copy.

Pass before Step 5: The skill is loaded (or its checklist is open in context). Do not classify severity or write findings until this gate clears.

Step 4: Load Skills

Read each applicable skill below (open its SKILL.md) so its guidance is in context before you review that area.

Always load (a non-trivial Remix v2 app exercises all six areas):

Detection telemetry (records what each area review will actually find; does not gate loading — all six skills always load, but record matches/non-matches for the final report's coverage section):

Condition Skill
app/routes/ present remix-v2-routing-review
loader / action exports found remix-v2-data-flow-review
\x3CForm>, useFetcher, useSubmit, useNavigation, native \x3Cform, or manual fetch( found remix-v2-forms-review
ErrorBoundary / useRouteError found remix-v2-error-boundaries-review
headers export, defer(, or .server.ts / .client.ts files or prefetch=/\x3CPrefetchPageLinks found remix-v2-perf-ssr-review
meta export or session/cookie helpers found remix-v2-meta-sessions-review

If a detection row returns no matches, record that explicitly in the report's coverage section — the skill still loads, but the reviewer should note the area had no surface to review.

Step 5: Review

If the agent supports subagents (and --parallel is requested or appropriate), dispatch one subagent per area in parallel; otherwise run the same areas sequentially in a single context. Both paths produce identical output.

Parallel path:

  1. Detect all areas upfront
  2. Dispatch one subagent per area, each loading its skill and reviewing its domain only
  3. Wait for all subagents to return
  4. Consolidate findings into a single output

Sequential path:

  1. Load applicable skills
  2. Review routing + nested route structure first
  3. Review loader/action data flow
  4. Review forms and progressive enhancement
  5. Review error boundaries
  6. Review meta, sessions, and headers/SSR
  7. Consolidate findings

Step 6: Verify Findings

Before reporting any issue:

  1. Re-read the actual code (not just diff context)
  2. For "unused" claims — did you search all references, including route-file conventions?
  3. For "missing" claims — did you check parent routes, root.tsx, or framework defaults?
  4. For syntax issues — did you verify against Remix v2 docs (not v1, not React Router v7)?
  5. Remove findings that are style preferences, not actual issues

Pass before promoting to Critical/Major: For that item, (2)–(4) are satisfied with a concrete artifact — opened file at FILE:LINE, grep output for references, or cited parent/framework code — not only diff context.

Step 7: Review Convergence

  • Single-pass: report ALL issues across routing, data flow, forms, errors, SSR/perf, meta/sessions, types, and security in one pass. Before submitting, ask whether your recommended fixes would themselves trigger new findings — if yes, include those now.
  • Scope: review only code in the diff and directly related existing code. Missing test coverage is ONE Minor issue, no implementation prescriptions. No new dependencies.
  • Fix complexity: real severity for fixes to existing code. Requests for net-new code (new modules, dependencies, test suites, extracted abstractions) are Informational, not blocking.
  • Re-review: verify previous fixes only — no new discovery, no re-flagging unfixed Minor/Nice-to-Have items.

Output Format

## Review Summary
[1-2 sentence overview]

## Issues

### Critical (Blocking)
1. [FILE:LINE] ISSUE_TITLE
   - Issue / Why / Fix

### Major (Should Fix)
2. [FILE:LINE] ISSUE_TITLE
   - Issue / Why / Fix

### Minor (Nice to Have)
N. [FILE:LINE] ISSUE_TITLE
   - Issue / Why / Fix

### Informational (For Awareness)
N. [FILE:LINE] SUGGESTION_TITLE
   - Suggestion / Rationale

## Good Patterns
- [FILE:LINE] Pattern description (preserve this)

## Verdict
Ready: Yes | No | With fixes 1-N (Critical/Major only; Minor items are acceptable)
Rationale: [1-2 sentences]

Post-Fix Verification

After fixes are applied, run:

npm run lint
npm run typecheck
npm run test

All checks must pass before approval.

Gates

Advance in order; do not skip a pass condition by restating it informally.

  1. Scope recordedPass when: You have the output of the Step 1 command (or an explicit substitute path list) naming what is in scope.
  2. Protocol + always skills loadedPass when: review-verification-protocol and all six remix-v2-*-review skills are loaded before the first severity judgment.
  3. Conditional skills auditedPass when: For each Step 2 detection row you either confirmed a match or recorded that the command returned no results.
  4. Critical/Major evidencePass when: Each such finding cites a FILE:LINE that exists in the tree and meets the Step 6 pass rule for that finding type.
  5. Single outputPass when: The Issues section uses one continuous numbering sequence and the deliverable satisfies Step 7 single-pass completeness.

Rules

  • Load skills BEFORE reviewing — not after.
  • Number every issue sequentially with FILE:LINE and clear Issue / Why / Fix.
  • This is Remix v2 — do not apply React Router v7, Remix v1, or Next.js ("use client", server components) conventions.
  • Run the Post-Fix Verification block after fixes; Verdict ignores Minor and Informational items.
安全使用建议
Install is reasonable for reviewing Remix v2 projects. Use it on repositories you trust, or run the verification commands inside a container/VM because npm scripts come from the project being reviewed.
能力评估
Purpose & Capability
The skill's stated purpose is Remix v2 code review, and its capabilities align with that: detect changed files, inspect Remix app structure, load related review guidance, optionally split review work by area, and report findings.
Instruction Scope
The instructions are scoped to the target repository and review workflow. Optional parallel subagents are disclosed and limited to review areas; no hidden role changes, exfiltration, or unrelated commands are present.
Install Mechanism
The artifact is a single non-executable SKILL.md with no package dependencies, installer scripts, or background setup.
Credentials
The skill uses ordinary local review commands such as git, grep, find, and post-fix npm verification. Running npm scripts can execute repository-defined code, so this should be done only where the project is trusted or sandboxed.
Persistence & Privilege
No persistence, privilege escalation, credential access, session-store access, network exfiltration, or long-running worker behavior is requested by the skill.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install review-remix-v2
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /review-remix-v2 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release: Comprehensive, parallelizable code review for Remix v2 projects. - Detects Remix v2 projects and analyzes all relevant technology areas automatically. - Loads and applies six detailed sub-skills for routing, data flow, forms, error boundaries, SSR/performance, and meta/sessions. - Can review in parallel using subagents or sequentially, producing identical structured output. - Enforces cross-skill verification protocol before issuing findings or severities. - Structured output highlights issues by severity, review coverage, and good patterns; enforces post-fix gates and verification steps.
元数据
Slug review-remix-v2
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Review Remix V2 是什么?

Comprehensive Remix v2 code review with per-area review skills, run in parallel where the agent supports subagents and sequentially otherwise. Detects Remix... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 15 次。

如何安装 Review Remix V2?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install review-remix-v2」即可一键安装,无需额外配置。

Review Remix V2 是免费的吗?

是的,Review Remix V2 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Review Remix V2 支持哪些平台?

Review Remix V2 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Review Remix V2?

由 Kevin Anderson(@anderskev)开发并维护,当前版本 v1.0.0。

💬 留言讨论