test-authoring

Memory

Search first: /mnemonic:search {relevant_keywords} Capture after: /mnemonic:capture {namespace} "{title}"

Run /mnemonic:list --namespaces to see available namespaces from loaded ontologies.

Test Authoring for Hook-Driven Frameworks

Write effective test definitions for the hook-driven automated test framework.

Mnemonic Integration

Before authoring tests, check mnemonic for test patterns:

# Search for test authoring memories
rg -i "test|expectation|contains|regex|depends_on" ~/.claude/mnemonic/ ./.claude/mnemonic/ --glob "*.memory.md" -l | head -5

# Check patterns namespace
rg -l "." ~/.claude/mnemonic/*/patterns/ --glob "*.memory.md" 2>/dev/null | xargs grep -l -i "test" 2>/dev/null | head -5

Apply recalled patterns:

• Effective expectation strategies
• Variable capture patterns that worked
• Test organization preferences
• Common ID naming conventions used

Test Definition Schema

Every test requires these fields:

Field	Required	Type	Purpose
id	Yes	string	Unique identifier (snake_case)
description	Yes	string	Human-readable explanation
action	Yes	string	Instruction for Claude to execute
expect	Yes	array	Validation expectations
category	No	string	Grouping for filtering
save_as	No	string	Variable name to capture output
depends_on	No	string	ID of prerequisite test
tags	No	array	Labels for filtering

Expectation Types

contains - Substring Match

Use for flexible text matching when exact format may vary.

expect:
  - contains: "successfully"
  - contains: "Memory created"

When to use:

• Success messages
• Partial output validation
• UI text that may have surrounding content

not_contains - Negative Match

Ensure unwanted content is absent.

expect:
  - not_contains: "error"
  - not_contains: "failed"
  - not_contains: "undefined"

When to use:

• Error detection
• Ensuring deprecated patterns removed
• Security validation (no sensitive data exposed)

regex - Pattern Match

Use for structured data with predictable patterns.

expect:
  - regex: "ID:\\s+([a-f0-9]{12})"
  - regex: "Count:\\s+\\d+"
  - regex: "(success|complete)"

When to use:

• ID extraction
• Numeric validation
• Multiple acceptable values
• Structured output parsing

Capture groups: Use () to capture values for save_as.

Variable Capture and Substitution

Capturing Values

Use save_as with a regex capture group:

- id: create_item
  action: "Create a new item named 'test'"
  expect:
    - regex: "ID:\\s+([a-f0-9]+)"
  save_as: item_id

The first capture group () value is stored in the variable.

Using Captured Values

Reference with ${variable_name}:

- id: retrieve_item
  action: "Retrieve item ${item_id}"
  depends_on: create_item
  expect:
    - contains: "test"

Dependency Chains

Build test sequences:

# Test 1: Create
- id: crud_create
  action: "Create memory with content 'test data'"
  expect:
    - regex: "ID:\\s+(\\w+)"
  save_as: memory_id

# Test 2: Read (depends on create)
- id: crud_read
  action: "Retrieve memory ${memory_id}"
  depends_on: crud_create
  expect:
    - contains: "test data"

# Test 3: Update (depends on read)
- id: crud_update
  action: "Update memory ${memory_id} with content 'updated'"
  depends_on: crud_read
  expect:
    - contains: "updated"

# Test 4: Delete (depends on update)
- id: crud_delete
  action: "Delete memory ${memory_id}"
  depends_on: crud_update
  expect:
    - contains: "deleted"
    - not_contains: "error"

Writing Effective Actions

Be Specific

Good:

action: "Call subcog_capture with content: 'Test pattern', namespace: 'patterns'"

Avoid:

action: "Create a memory"  # Too vague

Match Tool Names Exactly

For MCP tools, use the full tool name:

action: "Call mcp__plugin_subcog_subcog__subcog_capture with content: 'test'"

Include Required Parameters

action: "Call subcog_recall with query: 'authentication', limit: 5, mode: 'hybrid'"

Test Organization

Categories

Group tests by component or functionality:

Category	Use For
initialization	Setup and init tests
crud	Create, read, update, delete
search	Query and filter operations
validation	Input validation tests
error-handling	Error condition tests
integration	Cross-component tests
commands	Slash command tests
skills	Skill trigger tests
hooks	Hook behavior tests

Tags

Use tags for filtering test runs:

Tag	Meaning
smoke	Quick sanity checks
critical	Must-pass tests
regression	Bug fix verification
slow	Long-running tests
flaky	Known intermittent failures

ID Naming Convention

Pattern: category_component_action_variant

Examples:

• init_basic - Basic initialization
• crud_memory_create - Create memory
• search_semantic_empty - Semantic search with no results
• cmd_review_with_args - Review command with arguments

Common Patterns

Smoke Test Pattern

Quick validation that component works:

- id: component_smoke
  description: "Basic smoke test for component"
  category: smoke
  action: "Minimal invocation of component"
  expect:
    - not_contains: "error"
  tags: [smoke, critical]

CRUD Test Pattern

Full lifecycle test:

- id: entity_create
  action: "Create entity"
  save_as: entity_id
  tags: [crud]

- id: entity_read
  action: "Read entity ${entity_id}"
  depends_on: entity_create

- id: entity_update
  action: "Update entity ${entity_id}"
  depends_on: entity_read

- id: entity_delete
  action: "Delete entity ${entity_id}"
  depends_on: entity_update

Error Handling Pattern

Validate error responses:

- id: invalid_input_error
  description: "Verify error on invalid input"
  action: "Call tool with invalid_param: 'not-valid'"
  expect:
    - contains: "error"
    - regex: "(invalid|not found|failed)"
  tags: [error-handling]

Search Results Pattern

Validate search functionality:

- id: search_with_results
  action: "Search for 'known term'"
  expect:
    - contains: "results"
    - regex: "found\\s+\\d+"
    - not_contains: "no results"

Additional Resources

Reference Files

For detailed patterns and advanced techniques:

• references/expectation-patterns.md - Regex patterns for common outputs