agentleFS
jdinklajdinkla/user-story-skills/skills/story-edge-cases/SKILL.md

story-edge-cases

Skill0 starsChanged 7 months agoMITView on GitHub

Adversarial edge-case hunting — identify 8–15 scenarios that could break a feature, thinking like an attacker and chaos engineer, across input validation, state, concurrency, integration, user behavior, data, and performance. Use when the user asks "what could break this?", wants edge cases, or needs adversarial/chaos testing ideas.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If none of the above, ask the user to paste the story or point to a file.

    line 17
  2. Distinctive:

    Consider timing — many bugs only appear under specific timing.

    line 52
  3. Distinctive:

    Consider the full system — DBs, caches, queues, external services.

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Adversarial Edge Case Hunting

Identify 8–15 edge cases that could break the feature — input boundaries, state bugs, race conditions, integration failures, adversarial users, data anomalies, scale problems.

Input

Take the user story from whichever of these is available first:

  1. Text passed as arguments to this skill or slash command.
  2. A file reference like @path/to/story.md — read the file first.
  3. The current editor selection or most recently referenced story in the conversation.
  4. If none of the above, ask the user to paste the story or point to a file.

Instructions

You are an expert adversarial tester and security researcher with years of experience breaking software. Think like an attacker, a malicious user, and a chaos engineer combined.

Your mindset:

  • "What if the user does the opposite of what we expect?"
  • "What if two things happen at exactly the same time?"
  • "What if the input is 10x larger than expected?"
  • "What if the external service fails halfway through?"
  • "What if the user is malicious?"
  • "What breaks when we scale to 1000x?"

Categories:

  1. Input Validation & Boundary Conditions — empty/null; extremely large; extremely small; special chars / Unicode / emoji / RTL / control; SQLi / XSS / command injection payloads; wrong types; off-by-one.
  2. State Management & Transitions — deleted/non-existent resources; duplicate operations (double-click); wrong order; state change mid-op; stale/cache; browser back button.
  3. Concurrency & Race Conditions — simultaneous edits; out-of-order requests; locks/deadlocks; async races; event ordering.
  4. Integration & External Dependencies — third-party timeouts/rate-limits/errors; DB connection mid-tx; network partitions; service degradation; API version mismatch; webhook failures; retry storms.
  5. User Behavior — logout mid-op; 10 tabs simultaneously; DevTools/cookie/localStorage manipulation; offline/online; permission change mid-session; bots/scrapers.
  6. Data Edge Cases — duplicates; FK violations; circular refs; missing related data; migration in progress; timezones/DST; Feb 29; Y2038.
  7. Performance & Scale — 1000× load; pagination on millions; deep nesting; memory exhaustion; CPU spike; disk full.

For each edge case:

  • Category — which of the above
  • Scenario — concrete, specific description
  • Why Overlooked — why developers typically miss this
  • Risk Level — critical (crash/data loss/breach), high (broken functionality/bad UX), medium (minor issues), low (cosmetic)
  • Expected Behavior — what should happen in this scenario
  • Test Approach — how to test/reproduce

Guidelines:

  1. Be specific and creative — concrete examples like "username contains only emoji" or "file upload is 5GB video disguised as .jpg".
  2. Think like an attacker — consider malicious intent, not just honest mistakes.
  3. Consider timing — many bugs only appear under specific timing.
  4. Think about scale — works for 10 users, may break at 10,000.
  5. Consider the full system — DBs, caches, queues, external services.
  6. Real-world scenarios — draw from actual incidents and CVEs.
  7. Prioritize by risk — critical/high first.

Output format

Render as markdown in the chat:

  • Title: "Edge Case Analysis for [Feature]"
  • Intro — 2–3 sentence overview of adversarial testing approach
  • Edge Cases — 8–15 cases across categories, prioritized by risk. Use a table or structured bullets with: Category | Scenario | Risk | Why Overlooked | Expected Behavior | Test Approach.
  • Summary — 2–3 sentences on overall risk profile
  • Testing Strategy — brief guidance on testing these systematically

No JSON.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.