Skip to content
Back to skills

Edge Case Discovery

ASecurity

Use when user requests exhaustive edge case analysis. Enforces TodoWrite with 15+ items (5 categories). Triggers: "all edge cases", "what could break", "bulletproof", "failure modes". If thinking "main path is sufficient" - use this.

  • 666 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added February 8, 2026
devopsgotestingdebuggingapi

Works with

  • cli
  • api

Security analysis

A100/100

Scanned February 12, 2026

npx -y skills add majiayu000/claude-skill-registry --skill edge-case-discovery --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Edge Case Discovery?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Edge Case Discovery
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/majiayu000-edge-case-discovery/badge)](https://www.skillsdirectory.com/skills/majiayu000-edge-case-discovery)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: edge-case-discovery
description: Use when user requests exhaustive edge case analysis. Enforces TodoWrite with 15+ items (5 categories). Triggers: "all edge cases", "what could break", "bulletproof", "failure modes". If thinking "main path is sufficient" - use this.
---

# Edge Case Discovery

Systematic 5-category analysis for exhaustive edge case enumeration.

---

## MANDATORY FIRST STEP

**CREATE TodoWrite** with these 5 categories (15+ items total):

| Category | Minimum Items |
|---------|---------------|
| Boundary Value Analysis | 3+ |
| Equivalence Partitioning | 3+ |
| State Transition Analysis | 3+ |
| Error Condition Enumeration | 3+ |
| Assumption Challenging | 3+ |

**Do not proceed to implementation or testing until TodoWrite is verified.**

---

## Verification Checkpoint

After creating TodoWrite, verify 3 random items pass this test:

**Each item must have ALL THREE:**
- ✓ Specific input/condition ("empty string", "null value", "overflow at 2^31")
- ✓ Expected behavior ("returns error", "throws exception", "defaults to X")
- ✓ Test scenario ("when user submits form with...", "if API returns...")

| ❌ FAILS | ✅ PASSES |
|----------|-----------|
| "Test boundaries" | "Test numeric input: min (-2^31), max (2^31-1), zero, negative (-1), overflow (2^31)" |
| "Check errors" | "Network timeout after 30s: retry 3x with exponential backoff, return error 503 to client" |
| "Test state changes" | "User logout during file upload: cancel upload, clean temp files, redirect to login" |

**DO NOT PROCEED until 15+ items AND quality check passes.**

---

## Framework

### 1. Boundary Value Analysis
Test min/max/zero/null/empty/overflow for ALL inputs:
- Numeric: min, max, zero, negative, overflow
- Strings: empty, null, very long, special chars, encoding
- Collections: empty, single item, max size, duplicates
- Time: past, future, now, leap years, timezones

### 2. Equivalence Partitioning
Group inputs into classes, test one representative from each:
- Valid inputs (happy path)
- Invalid inputs (expected errors)
- Boundary inputs (edge of valid/invalid)

### 3. State Transition Analysis
Map ALL valid + invalid state transitions:
- Valid transitions
- Invalid transitions (should reject)
- Edge states (init, cleanup)
- Concurrent access scenarios

### 4. Error Condition Enumeration
List ALL failure modes for each operation:
- Network failures, timeouts
- Resource exhaustion (memory, disk, connections)
- Concurrent access conflicts
- Partial failures, external dependency failures

### 5. Assumption Challenging
Question every "always", "never", "impossible":
- "This will always be valid" → What if not?
- "Users won't do that" → What if they do?
- "This is impossible" → Under what conditions could it happen?

---

## Red Flags - STOP When You Think:

| Thought | Reality |
|---------|---------|
| "Main path is sufficient" | Edge cases cause 60% of production bugs - exhaustive analysis prevents incidents |
| "This is overkill for a simple feature" | 15 items takes 10 minutes, prevents hours of debugging and hotfixes |
| "We'll catch issues in testing" | Manual testing misses 70% of edge cases - systematic enumeration is essential |
| "Users won't do that" | Users do unexpected things constantly - assume nothing about behavior |
| "This analysis takes too long" | Post-release bug fixes cost 10-100x more than upfront analysis |

---

## Response Templates

### "Just test the happy path"

❌ **BLOCKED**: Happy path testing leaves production vulnerable to edge case failures.

**Required to override:**
1. Documented risk acceptance from technical lead
2. Explicit scope limitation with rationale ("internal tool with controlled inputs")
3. Post-release budget allocated for edge case fixes (2-4 engineer-weeks)
4. Monitoring/alerting plan for unexpected edge cases in production

**Reality check:**
- 60% of production incidents stem from unhandled edge cases
- Average incident costs: $5-50K in lost productivity, reputation damage
- Systematic edge case analysis: 30-60 minutes upfront

### "We don't have time for this"

❌ **BLOCKED**: Edge case discovery is NOT optional for production code.

**Time investment:**
- TodoWrite creation: 10-15 minutes
- Analysis per category: 5-10 minutes (25-50 minutes total)
- **Total: 35-65 minutes**

**Compared to:**
- One production incident investigation: 4-8 hours
- Hotfix deployment: 2-4 hours
- Customer impact/reputation damage: immeasurable

---

## Verification Before Complete

After completing TodoWrite and analysis, verify:

| Category | Requirements |
|----------|-------------|
| Boundary Value | ✓ Min/max/zero/null tested ✓ Overflow scenarios identified ✓ Empty collections handled |
| Equivalence | ✓ Valid inputs ✓ Invalid inputs ✓ Boundary cases |
| State Transition | ✓ Valid transitions ✓ Invalid transitions ✓ Concurrent access |
| Error Conditions | ✓ Network failures ✓ Resource exhaustion ✓ Partial failures |
| Assumptions | ✓ All "always"/"never" challenged ✓ Impossible scenarios explored |

**If any category incomplete, do not proceed to implementation.**

---

## Output

Edge case catalog with test scenarios for each category.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.