review

Analyze pull request diffs for SQL safety, LLM trust boundaries, and side effects.

Updated Mar 28, 2026

Search Tags:#security #llm #sql #review #diff #code-review #pull-request

Ready to run

npx skills add https://github.com/steadyfall/fspeek --skill review-steadyfall

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: review
Source: https://github.com/steadyfall/fspeek/tree/main/.claude/skills/gstack/review
Command: npx skills add https://github.com/steadyfall/fspeek --skill review-steadyfall

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Analyzes pull request diffs to surface SQL safety issues, LLM prompt-injection risks, and conditional side-effect problems before merging.

Core Features & Use Cases

Diff-based risk analysis for SQL safety, LLM trust boundaries, and side effects.
Proactive guidance for reviewers by surfacing likely risk areas in a PR.
Use Cases: during code reviews, pre-landing PR checks, or when asked to review a diff.

Quick Start

Provide a structured risk assessment for the current PR and recommended actions.

Frequently Asked Questions about review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I review a pull request diff for SQL injection and LLM prompt injection risks?▼

You can identify SQL safety issues in PR diffs by applying structured risk analysis that evaluates query construction and execution paths. This surfaces potential injection vulnerabilities and unsafe query patterns before the code lands.

What is the best way to check LLM trust boundaries during a code review?▼

Checking LLM trust boundaries during code review involves evaluating how user inputs interact with LLM prompts in the diff. This surfaces prompt-injection risks and unsafe data handoffs between system components to guide secure merging.

Can I use automated PR checks to surface conditional side effects in diffs?▼

Automated PR checks can surface conditional side effects by applying predefined prompts and structured analysis to the diff. This identifies hidden state mutations or unexpected behaviors triggered under specific conditions before landing.

How do I perform a pre-landing review for SQL safety and side effects?▼

Performing a pre-landing review for SQL safety and side effects requires evaluating the diff's query logic and conditional execution paths. This provides a structured risk assessment with recommended actions to prevent unsafe merges.

What types of code review risks does diff-based analysis identify?▼

Diff-based analysis identifies SQL safety issues, LLM prompt-injection vulnerabilities, and conditional side-effect problems. It evaluates trust boundaries and execution paths to proactively surface risks and guide reviewer decisions.

When do I need to run a structured risk assessment on a pull request?▼

You need to run a structured risk assessment on a pull request during code reviews, pre-landing checks, or when asked to review a diff. This is required when changes involve SQL queries, LLM interactions, or complex conditional logic.