review-edge-case-hunter

Walks every branching path in code or diffs and reports unhandled edge cases as JSON.

3|Updated Feb 26, 2026
One-click install
npx skills add https://github.com/stellar-zk/stellar-zk --skill review-edge-case-hunter-stellar-zk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review-edge-case-hunter
Source: https://github.com/stellar-zk/stellar-zk/tree/main/.claude/skills/review-edge-case-hunter
Command: npx skills add https://github.com/stellar-zk/stellar-zk --skill review-edge-case-hunter-stellar-zk

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Code reviews often miss boundary conditions and unguarded branches that only surface as production bugs. This Skill mechanically enumerates every control-flow path and domain boundary in a diff, file, or function, then reports only the paths lacking explicit handling. ## Core Features & Use Cases - Exhaustive Path Enumeration: Walks conditionals, loops, error handlers, early returns, and domain boundaries such as off-by-one loops, arithmetic overflow, null inputs, race conditions, and timeout gaps. - Diff-Aware Scoping: When given a diff, analyzes only changed hunks and boundaries directly reachable from changed lines; when given a full file or function, treats the entire content as scope. - Structured JSON Output: Returns findings as a strict JSON array with location, trigger condition, guard snippet, and potential consequence, with an empty array when nothing is unhandled. - Use Case: Paste a pull request diff before merging and receive a machine-readable list of every unguarded branch, such as a missing else clause on a new status check or an unbounded loop over user input. ## Quick Start Ask the AI to run an edge-case review on the attached diff or function and return only the unhandled paths as JSON.

Frequently Asked Questions about review-edge-case-hunter

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

FAQPage Schema
How do I find unhandled edge cases in a code diff?▼

Provide the diff as input and the review walks every branch reachable from the changed lines, checking for missing guards. It returns a JSON array listing each unhandled path with its location, trigger condition, a suggested guard snippet, and the potential consequence.

What kinds of edge cases does automated code review detect?▼

It derives edge classes from the content itself rather than a fixed checklist. Typical findings include missing else or default branches, unguarded null or empty inputs, off-by-one loops, arithmetic overflow, implicit type coercion, race conditions, and timeout gaps.

Can I review a full file instead of a diff for edge cases?▼

Yes. When no diff is provided, the entire file or function becomes the analysis scope and every branching path within it is walked. Content outside the provided input is ignored unless explicitly referenced.

What output format does the edge case review return?▼

It returns only a valid JSON array where each object has exactly four fields: location, trigger_condition, guard_snippet, and potential_consequence. An empty array is returned when no unhandled paths exist, with no extra prose or markdown.

What happens if the review input is empty or unreadable?▼

A halt condition triggers and the review returns a single JSON object reporting that the input was empty or undecodable, with a guard snippet asking for valid content. No further analysis steps are executed.