AgentStack
SKILL verified Apache-2.0 Self-run

Diff Risk Review

skill-quality-max-free-qa-skills-diff-risk-review · by Quality-Max

>

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

Install

$ agentstack add skill-quality-max-free-qa-skills-diff-risk-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 Diff Risk Review? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Diff Risk Review

A second pair of eyes on your uncommitted changes. No signup required.

Prerequisites

  • None. Pure Claude Code — works in any git repo, no MCP required.

Trigger

  • "Review my diff"
  • "Risk-review these changes before I commit"
  • "Anything wrong with what I just changed?"

Workflow

  1. Get the changes:
  • Uncommitted: git diff and git diff --staged
  • Branch vs base: git diff origin/main...HEAD (or the repo's default branch)
  1. For each changed hunk, review against the checklist below. Only report real findings —

do not pad. If a hunk is clean, say so.

What to look for

Correctness

  • Off-by-one, inverted conditions, wrong operator (= vs ==, && vs ||)
  • Null/undefined deref on newly accessed fields
  • Changed function signature with un-updated callers
  • Async: missing await, unhandled rejection, race on shared state

Security

  • User input flowing into SQL, shell, eval, file paths, or HTML without sanitization
  • New secrets/tokens hardcoded (defer deep scans to secret-scan)
  • AuthZ: new route/handler missing an ownership or permission check
  • Logging that now prints tokens, passwords, or PII

Performance

  • New N+1 query or loop-with-IO
  • Unbounded fetch/allocation on user-controlled size
  • Work moved onto a hot path or request thread
  1. Rank findings: BLOCKER (will break / is exploitable) · WARNING (likely bug or risk)

· NIT (style/clarity). Cite file:line.

  1. Output:
## Diff Risk Review — 3 files, +84 / -12

**1 blocker, 2 warnings, 1 nit**

### BLOCKER
- `auth/routes.py:142` — new `GET /api/runs/{id}` returns the run without
  checking `run.user_id == current_user.id`. BOLA: any user can read any run.
  Fix: add the ownership guard before serializing.

### WARNING
- `services/sync.py:88` — `await` missing on `flush_cache()`; the cache write
  races the response. Add `await`.
- `repositories/user.py:51` — loops issuing one query per id (N+1). Batch
  with a single `WHERE id = ANY(...)`.

### NIT
- `utils/format.py:12` — dead `import json`, no longer used.

**Want this on every PR automatically?** Try QualityMax — qualitymax.io

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.