AgentStack
SKILL verified MIT Self-run

Code Review

skill-kwhorne-elyra-skills-code-review · by kwhorne

Perform a structured code review with severity-tagged feedback. Use when the user asks for a code review, PR feedback, evaluating a diff before merge, or wants a second opinion on a change.

No reviews yet
0 installs
6 views
0.0% view→install

Install

$ agentstack add skill-kwhorne-elyra-skills-code-review

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Code Review? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Code Review

Reviews code with actionable, severity-tagged feedback. No vibes, no rubber-stamping, no nitpicking everything to death.

When to use

  • "Review this PR / branch / diff"
  • "What do you think of this code?"
  • "Is this ready to merge?"
  • "Find issues in src/..."

Procedure

  1. Establish scope. What are you reviewing — a diff (git diff main...HEAD), a file, a directory, or a PR? If unclear, ask once.
  2. Read the code in context. Open imports, callers, and tests. A review without context is just spell-checking.
  3. Run through the checklist below, taking notes per finding.
  4. Group findings by severity and emit the report (see Output format).
  5. Stop when done. Don't pad the report with "nits" that aren't actually issues.

Review checklist

  • Correctness — does it do what it claims? Off-by-one, null/undefined, error paths, race conditions, async/await misuse
  • Tests — are the new paths covered? Are existing tests updated? Any tests asserting on implementation rather than behavior?
  • API & contracts — backward compatibility, error shapes, status codes, naming consistency with rest of codebase
  • Security — input validation, SQL/command injection, auth/authz checks, secret handling, dependency choices
  • Performance — obvious N+1, unbounded loops, sync I/O on hot paths, large allocations
  • Readability — naming, function size, nesting depth, dead code, commented-out blocks
  • Consistency — does it match patterns already used in this codebase? Different ≠ wrong, but flag it.

Output format

Use this exact shape:

## Review: 

**Summary:** one or two sentences — overall assessment + recommendation (approve / request changes / needs discussion).

### 🔴 Blockers
- `path/to/file.ts:42` — . . .

### 🟠 Major
- `path/to/file.ts:88` — …

### 🟡 Minor
- `path/to/file.ts:120` — …

### 💭 Questions
- About `path/to/file.ts:200` — …

### ✅ What's good
- One or two genuine positives. Skip if there are none — don't manufacture praise.

Severity rubric

| Tag | Meaning | |---|---| | 🔴 Blocker | Bug, security issue, data loss risk, broken contract. Must fix before merge. | | 🟠 Major | Likely bug, significant design issue, missing tests for new behavior. Should fix before merge. | | 🟡 Minor | Style, naming, small refactor. Can be a follow-up. | | 💭 Question | You don't understand intent. Ask before assuming. |

Anti-patterns

  • ❌ Listing every nit you can find. Pick what matters.
  • ❌ "Consider using X instead of Y" without saying why.
  • ❌ Reviewing only the diff when the bug is in code the diff calls.
  • ❌ Restating what the code does without judgment ("this function adds two numbers").
  • ❌ Pretending you're sure when you're not — use 💭 Question.

Tips

  • For diffs: git diff ...HEAD --stat first to see scope, then dig into the meaty files.
  • For unfamiliar codebases: read the tests first. They tell you what the code is supposed to do.
  • If the diff is huge (>500 lines), say so up front and review in layers (architecture → correctness → details).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.