Install
$ agentstack add skill-kumaran-is-claude-code-onboarding-code-reviewer ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo 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.
About
Iron Law: Never approve code without reading the actual implementation; always provide severity-bucketed findings with file:line evidence.
Code Reviewer
General-purpose code review skill covering security, quality, performance, and best practices.
When to Use
- After writing or modifying code
- During PR reviews
- When auditing code quality
Process
- Identify changed files via
git diffor user request - Read [reference/code-review-checklist.md](reference/code-review-checklist.md) for review categories, severity levels, and output format
- Review each file against the checklist
- Report findings by severity (Critical > High > Medium > Low)
Reference Files
| File | Contents | Load When | |------|----------|-----------| | reference/code-review-checklist.md | Security checks, code quality, performance, best practices, output format | Reviewing any code change, pre-PR checklist, security audit |
Error Handling
If no changes are found, report "No changes detected" and list the files/paths searched. If a referenced file cannot be read, report the missing file and continue with available context.
Anti-Patterns
- Never rubber-stamp without file:line — "looks good" with no evidence is a failed review
- Never mark APPROVED if any Critical or High finding is unresolved — severity discipline is non-negotiable
- Never review from memory — always read the actual changed files; filenames do not tell you what changed
- Never scope-creep the review — report what changed, not what you wish were different in adjacent code
- Never merge findings across files — report per-file with specific line numbers, not aggregate impressions
- Avoid vague findings — "this could be a security issue" is not a finding; "SQL built via string concatenation at service.ts:47 allows injection" is
Verify
After completing a review:
# Confirm all files you reviewed are the ones that actually changed
git diff --name-only HEAD~1
# Confirm no new linting errors introduced
npm run lint 2>&1 | grep -E "error|warning" | head -20
# Confirm tests still pass after changes
npm test -- --passWithNoTests 2>&1 | tail -5
Report format after verification:
REVIEW COMPLETE:
- Files reviewed: N (list them)
- Findings: X Critical, Y High, Z Medium, W Low
- Verdict: APPROVED / NEEDS_REVIEW / REJECT
- Evidence: [file:line for each finding]
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: kumaran-is
- Source: kumaran-is/claude-code-onboarding
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.