Dev Fuzz Testing (Gate 4)

Overview

Ensure critical parsing and input handling code has fuzz tests to discover crashes and edge cases through random input generation.

Core principle: Fuzz tests find bugs you didn't think to test for. They're mandatory for all code that handles external input.

- No fuzz functions = FAIL - Seed corpus < 5 entries = FAIL - Any crash found = FAIL (fix and re-run) </block_condition>

CRITICAL: Role Clarification

This skill ORCHESTRATES. QA Analyst Agent (fuzz mode) EXECUTES.

Who	Responsibility
This Skill	Gather requirements, dispatch agent, track iterations
QA Analyst Agent	Write fuzz tests, generate corpus, run fuzz

---

Standards Reference

MANDATORY: Load testing-fuzz.md standards via WebFetch.

https://raw.githubusercontent.com/LerianStudio/ring/main/dev-team/docs/standards/golang/testing-fuzz.md </fetch_required>

---

Step 1: Validate Input

REQUIRED INPUT:
- unit_id: [task/subtask being tested]
- implementation_files: [files from Gate 0]
- language: [go only for native fuzz]

OPTIONAL INPUT:
- gate3_handoff: [full Gate 3 output]

if any REQUIRED input is missing:
  → STOP and report: "Missing required input: [field]"

if language != "go":
  → STOP and report: "Native fuzz testing only supported for Go (Go 1.18+)"

Step 2: Dispatch QA Analyst Agent (Fuzz Mode)

Task tool:
  subagent_type: "ring:qa-analyst"
  model: "opus"
  prompt: |
    **MODE:** FUZZ TESTING (Gate 4)

    **Standards:** Load testing-fuzz.md

    **Input:**
    - Unit ID: {unit_id}
    - Implementation Files: {implementation_files}
    - Language: {language}

    **Requirements:**
    1. Create fuzz functions (FuzzXxx naming)
    2. Add seed corpus (minimum 5 entries per function)
    3. Run fuzz tests for 30 seconds
    4. Report any crashes found

    **Output Sections Required:**
    - ## Fuzz Testing Summary
    - ## Corpus Report
    - ## Handoff to Next Gate

Step 3: Evaluate Results

Parse agent output:

if "Status: PASS" in output:
  → Gate 4 PASSED
  → Return success with metrics

if "Status: FAIL" in output:
  → Dispatch fix to implementation agent
  → Re-run fuzz tests (max 3 iterations)
  → If still failing: ESCALATE to user

Step 4: Generate Output

## Fuzz Testing Summary
**Status:** {PASS|FAIL}
**Fuzz Functions:** {count}
**Corpus Entries:** {count}
**Crashes Found:** {count}

## Corpus Report
| Function | Entries | Crashes |
|----------|---------|---------|
| {function_name} | {count} | {count} |

## Handoff to Next Gate
- Ready for Gate 5 (Property Testing): {YES|NO}
- Iterations: {count}

---

Anti-Rationalization Table

Rationalization	Why It's WRONG	Required Action
"Unit tests cover edge cases"	You can't test what you don't think of. Fuzz finds unknowns.	Write fuzz tests
"Code is too simple for fuzz"	Simple code can still crash on malformed input.	Write fuzz tests
"Fuzz testing is slow"	30 seconds per function. Crashes in production are slower.	Write fuzz tests
"We validate input anyway"	Validation can have bugs too. Fuzz tests the validators.	Write fuzz tests

---