arinhub-code-reviewer

Original🇺🇸 English
Translated

Use this skill to review code with prefix 'ah review'. Use when asked to 'ah review PR 123', 'ah review my changes', 'ah review pr

1installs
Added on

NPX Install

npx skill4agent add arinhubcom/arinhub arinhub-code-reviewer

Code Reviewer

Orchestrate a comprehensive code review by running multiple review strategies in parallel, merging and deduplicating findings into a review file. Supports both remote PRs and local uncommitted changes.

Input

  • PR number or URL (optional): Accepts
    123
    ,
    #123
    , or full URL. If omitted, reviews local changes.

Procedure

1. Determine Review Target

  • Remote PR: If the user provides a PR number or URL (e.g., "Review PR #123"), target that remote PR. Set
    MODE=remote
    .
  • Local Changes: If no specific PR is mentioned, or if the user asks to "review my changes", target the current local file system changes (staged and unstaged). Set
    MODE=local
    .

2. Resolve Identifier and Repository

If
MODE=remote
:
Extract the PR number. Determine the repository name from git remote or the provided URL.
PR_NUMBER=<extracted number>
REPO_NAME=<repository name, e.g. "my-app">
REVIEW_FILE=~/.agents/arinhub/code-reviews/pr-code-review-${REPO_NAME}-${PR_NUMBER}.md
If
MODE=local
:
Determine the repository name from git remote. Use the current branch name for identification, sanitizing slashes to dashes so file paths remain valid.
REPO_NAME=<repository name>
BRANCH_NAME=$(git branch --show-current | tr '/' '-')
REVIEW_FILE=~/.agents/arinhub/code-reviews/local-code-review-${REPO_NAME}-${BRANCH_NAME}.md
Create
~/.agents/arinhub/code-reviews/
and
~/.agents/arinhub/diffs/
directories if they do not exist.

3. Initialize Review File

If
MODE=remote
:
Create the review file with a header:
markdown
# PR Review: ${REPO_NAME} #${PR_NUMBER}

**Date:** <current date>
**PR:** <PR URL>

## Issues

<!-- Issues from parallel review agents merged below. No duplicates. -->
If
MODE=local
:
Create the review file with a header:
markdown
# Local Review: ${REPO_NAME} (${BRANCH_NAME})

**Date:** <current date>
**Branch:** ${BRANCH_NAME}

## Issues

<!-- Issues from parallel review agents merged below. No duplicates. -->

4. Prepare Diff and Working Tree

Save the diff to a shared file so subagents can read it. In remote mode, also check out the PR branch so tools that require a working tree (e.g.,
react-doctor
) operate on the correct code.
If
MODE=remote
:
bash
DIFF_FILE=~/.agents/arinhub/diffs/pr-diff-${REPO_NAME}-${PR_NUMBER}.diff

# Save the current branch so we can return to it after the review.
ORIGINAL_BRANCH=$(git branch --show-current)

# Stash any uncommitted local changes to prevent data loss during checkout.
git stash --include-untracked -m "arinhub-code-reviewer: auto-stash before PR checkout"

gh pr diff ${PR_NUMBER} > ${DIFF_FILE}

# Check out the PR branch to ensure the working tree reflects the PR code for subagents that require it (e.g., react-doctor).
gh pr checkout ${PR_NUMBER}
If
MODE=local
:
bash
DIFF_FILE=~/.agents/arinhub/diffs/local-diff-${REPO_NAME}-${BRANCH_NAME}.diff
git diff HEAD > "${DIFF_FILE}"
No checkout is needed in local mode — the working tree already contains the changes.

5. Detect React Code

Spawn a subagent to analyze
${DIFF_FILE}
and determine whether the changes contain React code. The subagent must read the diff file and return
HAS_REACT=true
or
HAS_REACT=false
.
Set
HAS_REACT=true
if any of these conditions are found in the diff:
  • File extensions: Changed files include
    .tsx
    ,
    .jsx
    , or paths under common React directories (e.g.,
    components/
    ,
    hooks/
    ,
    pages/
    )
  • React core imports:
    import ... from 'react'
    ,
    import ... from "react"
    ,
    require('react')
    ,
    require("react")
  • React DOM:
    import ... from 'react-dom'
    ,
    import ... from 'react-dom/client'
  • JSX syntax: Diff hunks contain JSX elements (
    <Component
    ,
    <div
    ,
    />
    ,
    React.createElement
    )
  • React hooks: Usage of
    useState
    ,
    useEffect
    ,
    useRef
    ,
    useMemo
    ,
    useCallback
    ,
    useContext
    ,
    useReducer
    ,
    useLayoutEffect
    , or custom
    use*
    hooks
  • React ecosystem packages: Imports from
    react-router
    ,
    react-hook-form
    ,
    @tanstack/react-query
    ,
    @tanstack/react-table
    ,
    react-redux
    ,
    zustand
    ,
    jotai
    ,
    recoil
    ,
    next
    ,
    @next/
    ,
    styled-components
    ,
    @emotion/
    , or similar React-centric libraries
Otherwise set
HAS_REACT=false
.

6. Launch Parallel Review Subagents

Spawn subagents in parallel (do not wait for one to finish before starting the next). Subagents A, B, and C must return ONLY a structured list of issues using the format from the Issue Format Reference section below. Subagent D (if launched) must return both a structured list of issues and the full
react-doctor
diagnostic report (the report is appended separately in Step 9). No subagent may submit a review — they only return findings.
Pass the diff file path (
${DIFF_FILE}
) to each subagent so they can read the diff directly. Inform each subagent that the working tree is already on the correct branch (PR branch in remote mode, current branch in local mode). No subagent should run
gh pr checkout
or switch branches.
  • If
    HAS_REACT=true
    : spawn four subagents (A, B, C, D).
  • If
    HAS_REACT=false
    : spawn three subagents (A, B, C) — skip Subagent D.

Subagent A: code-reviewer

If
MODE=remote
:
Spawn a subagent to review PR
${PR_NUMBER}
using the
code-reviewer
skill. Pass
${DIFF_FILE}
for diff context. If
MODE=local
:
Spawn a subagent to review local changes using the
code-reviewer
skill. Pass
${DIFF_FILE}
for diff context.
Instruct it to return only the list of issues found — no review submission.

Subagent B: octocode-roast

Spawn a subagent to invoke the
octocode-roast
skill. Pass
${DIFF_FILE}
so it can read the diff directly.
If
MODE=remote
:
Instruct it to review the diff in
${DIFF_FILE}
for PR
${PR_NUMBER}
. If
MODE=local
:
Instruct it to review the diff in
${DIFF_FILE}
for local changes.
Instruct it to return only the list of issues found — no review submission.

Subagent C: pr-review-toolkit

If
MODE=remote
:
Spawn a subagent to review PR
${PR_NUMBER}
using the
pr-review-toolkit:review-pr
command with
all parallel
mode. Pass
${DIFF_FILE}
for diff context. If
MODE=local
:
Spawn a subagent to review local changes using the
pr-review-toolkit:review-pr
command with
all parallel
mode. Pass
${DIFF_FILE}
for diff context.
Instruct it to return only the list of issues found — no review submission.

Subagent D: react-doctor (only if
HAS_REACT=true
)

Spawn a subagent to run
react-doctor
on the working tree. The tool runs via
npx -y react-doctor@latest . --verbose --diff
and requires the working tree to be on the correct branch (already ensured by Step 4).
If
MODE=remote
:
Inform the subagent that the PR branch is already checked out. Instruct it to review the React code in the current working tree with diff context from
${DIFF_FILE}
. If
MODE=local
:
Inform the subagent that the working tree already contains the local changes. Instruct it to review the React code in the current working tree with diff context from
${DIFF_FILE}
.
Instruct it to diagnose React-specific issues (performance, hooks misuse, component anti-patterns, security) and return both the structured list of issues (using the Issue Format Reference) and the full
react-doctor
diagnostic report. No review submission.

7. Merge and Deduplicate Issues

Collect issues from all subagents (three or four, depending on
HAS_REACT
) and deduplicate:
  1. Parse each subagent's response into individual issues.
  2. For each issue, create a fingerprint from:
    file path
    +
    line number range
    +
    concern category
    .
  3. Two issues are duplicates if they share the same file, overlapping line ranges (within ±5 lines), and address the same concern (use semantic comparison, not exact string matching).
  4. When duplicates are found, keep the most detailed/actionable version.
  5. Tag each kept issue with its source(s):
    [code-reviewer]
    ,
    [octocode-roast]
    ,
    [pr-review-toolkit]
    ,
    [react-doctor]
    , or combination if multiple agents found it.

8. Write Issues to Review File

Append deduplicated issues to the review file, grouped by severity:
markdown
### Critical

- **[source]** `path/to/file.ts:42` — Description of the issue.

  ```ts
  // the problematic code from the PR diff
  const result = unsafeOperation(input);
  ```

  ```diff
  - const result = unsafeOperation(input);
  + const result = safeOperation(sanitize(input));
  ```

### Improvements

- **[source]** `path/to/file.ts:88-95` — Description of the issue.

  ```ts
  // the problematic code from the PR diff
  items.forEach((item) => {
    process(item);
  });
  ```

  ```diff
  - items.forEach(item => {
  -   process(item);
  - });
  + await Promise.all(items.map(item => process(item)));
  ```

### Nitpicks

- **[source]** `path/to/file.ts:12` — Description of the issue.

  ```ts
  // the relevant code snippet
  let x = getValue();
  ```

---

**Total issues:** N (X critical, Y improvements, Z nitpicks)
**Sources:** code-reviewer, octocode-roast, pr-review-toolkit[, react-doctor] (include react-doctor only if HAS_REACT=true)

9. React Health Report (only if
HAS_REACT=true
)

If
HAS_REACT=true
, append the full output from the
react-doctor
subagent (Subagent D) to the review file under a dedicated section:
markdown
## React Health

<full react-doctor report>
This section captures React-specific diagnostics (performance, hooks, component patterns, security) separately from the general deduplicated issues above.
If
HAS_REACT=false
, skip this section entirely.

10. Verify Requirements Coverage

Spawn a subagent to verify requirements coverage using the
arinhub-verify-requirements-coverage
skill. Pass the diff file path (
${DIFF_FILE}
) so the subagent can read the diff directly without fetching it again. The subagent must return the full requirements coverage report in markdown format.
If
MODE=remote
:
Pass PR
${PR_NUMBER}
and
${DIFF_FILE}
to the subagent. It will use the diff file for analysis and resolve the linked issue automatically.
If
MODE=local
:
Pass
${DIFF_FILE}
to the subagent. The subagent will attempt to extract the linked issue number from the branch name (e.g.,
feature/42-description
,
fix/42
,
issue-42-description
). If no issue can be determined, the subagent will skip coverage verification and report that no linked issue was found.
Append the returned coverage report to the end of the review file under a new section:
markdown
## Requirements Coverage

<coverage report content from arinhub-verify-requirements-coverage>

11. Submit PR Review (only if
MODE=remote
)

Skip this step if
MODE=local
.
Spawn a subagent to submit the review for PR
${PR_NUMBER}
using the
arinhub-submit-code-review
skill. Pass the review file path (
${REVIEW_FILE}
) so the subagent reads issues from it. The subagent must follow the
arinhub-submit-code-review
procedure for deduplication against existing PR comments before submission.

12. Restore Working Tree (only if
MODE=remote
)

Skip this step if
MODE=local
.
Return to the original branch and restore any stashed changes from Step 4:
bash
git checkout ${ORIGINAL_BRANCH}

# Restore stashed changes if the stash was created in Step 4.
git stash list | grep -q "arinhub-code-reviewer: auto-stash" && git stash pop

13. Report to User

If
MODE=remote
:
Present a summary:
  • Path to the review file
  • Total issues found (by severity)
  • PR coverage percentage
  • Whether the review was submitted successfully
  • The PR URL for reference
If
MODE=local
:
Present the review file (
${REVIEW_FILE}
) content to the user and a summary:
  • Path to the review file
  • Total issues found (by severity)
  • Requirements coverage percentage (if available)
  • Branch name and list of changed files reviewed

Issue Format Reference

Each issue in a subagent response must follow this structure:
markdown
- **Severity:** critical | improvement | nitpick
  **File:** path/to/file.ts
  **Line(s):** 42 (or 42-50)
  **Description:** Clear explanation of the problem.
  **Code:**
  ```ts
  // the problematic code from the PR diff
  const result = unsafeOperation(input);
  ```
  **Suggestion:**
  ```diff
  - const result = unsafeOperation(input);
  + const result = safeOperation(sanitize(input));
  ```

Important Notes

  • Review subagents run in parallel to minimize total review time (three or four, depending on whether the changes contain React code).
  • The
    react-doctor
    subagent is only launched when the diff contains
    .tsx
    /
    .jsx
    files or React imports. This avoids unnecessary React diagnostics on non-React changes.
  • The review file is the single source of truth — all findings are merged there before submission.
  • Deduplication uses semantic comparison: if two agents flag the same concern on the same code, only one entry is kept.
  • The review file persists at
    ~/.agents/arinhub/code-reviews/
    for future reference and audit.
  • If a subagent fails or times out, proceed with results from the remaining agents and note the failure in the review file.
  • The diff file persists at
    ~/.agents/arinhub/diffs/
    and is shared read-only across all subagents. The PR branch checkout happens once in Step 4 before subagents launch — no subagent should run
    gh pr checkout
    or switch branches on its own.
  • In
    MODE=local
    , step 11 (Submit PR Review) is skipped — the review is output only to the review file and presented to the user. Step 10 (Verify Requirements Coverage) runs if a linked issue can be determined from the branch name or user input.