Doc Spec

doc-spec

Purpose

Create Technical Specifications (SPEC) - Layer 10 artifact in the SDD workflow that defines implementation-ready specifications in YAML format, providing complete technical details for code generation.

Layer: 10

Upstream: BRD (Layer 1), PRD (Layer 2), EARS (Layer 3), BDD (Layer 4), ADR (Layer 5), SYS (Layer 6), REQ (Layer 7), IMPL (Layer 8), CTR (Layer 9)

Downstream Artifacts: TASKS (Layer 11), IPLAN (Layer 12), Code (Layer 13)

Prerequisites

Upstream Artifact Verification (CRITICAL)

Before creating this document, you MUST:

1. List existing upstream artifacts:

   ls docs/BRD/ docs/PRD/ docs/EARS/ docs/BDD/ docs/ADR/ docs/SYS/ docs/REQ/ docs/IMPL/ docs/CTR/ 2>/dev/null
   

2. Reference only existing documents in traceability tags

3. Use null only when upstream artifact type genuinely doesn't exist

4. NEVER use placeholders like BRD-XXX or TBD

5. Do NOT create missing upstream artifacts - skip functionality instead

Before creating SPEC, read:

1. Shared Standards: .claude/skills/doc-flow/SHARED_CONTENT.md
2. Upstream REQ: Read atomic requirements (PRIMARY SOURCE)
3. Upstream CTR: Read contracts if Layer 9 created
4. Template: ai_dev_flow/SPEC/SPEC-TEMPLATE.yaml
5. Creation Rules: ai_dev_flow/SPEC/SPEC_CREATION_RULES.md
6. Validation Rules: ai_dev_flow/SPEC/SPEC_VALIDATION_RULES.md
7. Validation Script: ./ai_dev_flow/scripts/validate_spec_template.sh (under development - use template for manual validation until available)

Reserved ID Exemption (SPEC-00_*)

Scope: Documents with reserved ID 000 are FULLY EXEMPT from validation.

Pattern: SPEC-00_*.md, SPEC-00_*.yaml

Document Types:

• Index documents (SPEC-00_index.md)
• Traceability matrix templates (SPEC-00_TRACEABILITY_MATRIX-TEMPLATE.md)
• Glossaries, registries, checklists

Rationale: Reserved ID 000 documents are framework infrastructure (indexes, templates, reference materials), not project artifacts requiring traceability or quality gates.

Validation Behavior: Skip all checks when filename matches SPEC-00_* pattern.

When to Use This Skill

Use doc-spec when:

• Have completed BRD through REQ (Layers 1-7)
• Ready to create implementation-ready specifications
• Preparing for code generation or implementation
• Need complete technical details in structured format
• You are at Layer 10 of the SDD workflow

SPEC-Specific Guidance

1. YAML Format (MANDATORY)

Format: Pure YAML (not markdown)

Structure:

# SPEC-NN: [Specification Title]

metadata:
  spec_id: SPEC-01
  title: "Order Validation Service Specification"
  version: "1.0.0"
  created_date: "2025-01-15"
  updated_date: "2025-01-15"
  status: "approved"
  owner: "team-backend"
  task_ready_score: "✅ 95% (Target: ≥90%)"

cumulative_tags:
  brd: ["BRD.01.01.03"]
  prd: ["PRD.01.07.02"]
  ears: ["EARS.01.25.01"]
  bdd: ["BDD.01.14.01"]
  adr: ["ADR-033", "ADR-045"]
  sys: ["SYS.01.26.01"]
  req: ["REQ.01.27.01"]
  impl: ["IMPL.01.29.01"]  # optional
  contracts: ["CTR-01"]  # optional

overview:
  purpose: "Define trade order validation service implementation"
  scope: "Validate trade orders against position limits and business rules"
  requirements:
    - "REQ-risk-limits-01"
    - "REQ-risk-limits-02"

architecture:
  pattern: "layered"
  layers:
    - name: "controller"
      technology: "FastAPI"
      description: "REST API endpoint handlers"
    - name: "service"
      technology: "Python"
      description: "Business logic and validation"
    - name: "repository"
      technology: "SQLAlchemy"
      description: "Database access layer"

interfaces:
  api_endpoints:
    - endpoint: "/api/v1/trades/validate"
      method: "POST"
      contract_ref: "CTR-01"
      authentication: "Bearer token"
      rate_limit: "@threshold: PRD.NN.limit.api.requests_per_second"
      rate_limit_window: "1min"

  data_models:
    - model: "TradeOrderRequest"
      schema_ref: "CTR-01#/components/schemas/TradeOrderRequest"
    - model: "ValidationResponse"
      schema_ref: "CTR-01#/components/schemas/ValidationResponse"

implementation:
  modules:
    - name: "controllers/trade_validation_controller.py"
      purpose: "API endpoint handlers"
      dependencies: ["services/trade_validator.py"]

    - name: "services/trade_validator.py"
      purpose: "Business logic and validation"
      dependencies:
        - "repositories/position_repository.py"
        - "models/trade_order.py"

    - name: "repositories/position_repository.py"
      purpose: "Database access for positions"
      dependencies: ["database/connection.py"]

  functions:
    - name: "validate_trade_order"
      module: "services/trade_validator.py"
      signature: "async def validate_trade_order(order: TradeOrderRequest) -> ValidationResponse"
      purpose: "Validate trade order against all rules"
      algorithm:
        - "1. Validate symbol exists"
        - "2. Check quantity is positive"
        - "3. Validate price within range"
        - "4. Check position limits"
        - "5. Return validation result"

error_handling:
  error_codes:
    - code: "INVALID_SYMBOL"
      http_status: 400
      message: "Symbol not found in approved list"
      recovery: "user_correction"

    - code: "LIMIT_EXCEEDED"
      http_status: 403
      message: "Position limit exceeded"
      recovery: "reduce_position"

configuration:
  environment_variables:
    - name: "MAX_POSITION_DELTA"
      type: "float"
      default: "0.50"
      required: true

  feature_flags:
    - name: "enable_strict_validation"
      default: false
      description: "Enable enhanced validation rules"

testing:
  unit_tests:
    - test: "test_validate_valid_order"
      module: "tests/unit/test_trade_validator.py"
      coverage_target: 95

  integration_tests:
    - test: "test_validation_endpoint"
      module: "tests/integration/test_trade_api.py"

  performance_tests:
    - test: "test_validation_latency"
      target: "P95 < 50ms"

deployment:
  container:
    image: "trade-validator:1.0.0"
    base: "python:3.11-slim"

  resources:
    cpu: "1000m"
    memory: "512Mi"

  scaling:
    min_replicas: 2
    max_replicas: 10
    target_cpu: 70

monitoring:
  metrics:
    - name: "validation_latency_ms"
      type: "histogram"
      labels: ["endpoint", "status"]

    - name: "validation_errors_total"
      type: "counter"
      labels: ["error_code"]

  alerts:
    - alert: "HighValidationLatency"
      condition: "P95 > 100ms"
      severity: "warning"

traceability:
  upstream_sources:
    - artifact: "BRD-01"
      sections: ["section-3"]
    - artifact: "PRD-01"
      sections: ["feature-2"]
    - artifact: "REQ-risk-limits-01"
      sections: ["all"]

  downstream_artifacts:
    - "TASKS-01"
    - "IPLAN-01"
    - "Code: src/services/trade_validator.py"

2. Element ID Format (MANDATORY)

Pattern: SPEC.{DOC_NUM}.{ELEM_TYPE}.{SEQ} (4 segments, dot-separated)

REMOVED PATTERNS - Do NOT use legacy formats:

• STEP-XXX - Use SPEC.NN.15.SS instead
• IF-XXX or INT-XXX - Use SPEC.NN.16.SS instead
• DM-XXX or MODEL-XXX - Use SPEC.NN.17.SS instead
• VR-XXX - Use SPEC.NN.21.SS instead

Reference: ID_NAMING_STANDARDS.md - Cross-Reference Link Format

3. Required Top-Level Sections

MANDATORY Sections:

1. metadata: Spec ID, title, version, dates, status, owner, task_ready_score
2. cumulative_tags: All upstream tags (7-9 tags depending on layers)
3. overview: Purpose, scope, requirements
4. architecture: Pattern, layers, technologies
5. interfaces: API endpoints, data models
6. implementation: Modules, functions, algorithms
7. error_handling: Error codes, HTTP status, recovery
8. configuration: Environment variables, feature flags
9. testing: Unit, integration, performance tests
10. deployment: Container, resources, scaling
11. monitoring: Metrics, alerts, logging
12. traceability: Upstream sources, downstream artifacts

4. TASKS-Ready Scoring System

Purpose: Measures SPEC maturity and readiness for progression to TASKS implementation planning.

Format in Metadata:

metadata:
  task_ready_score: "✅ 95% (Target: ≥90%)"

Status and TASKS-Ready Score Mapping:

Scoring Criteria:

• YAML Completeness (25%): All metadata fields, traceability chain, all sections populated
• Interface Definitions (25%): External APIs with CTR contracts, internal interfaces, data schemas
• Implementation Specifications (25%): Behavior enables code generation, performance/security quantifiable
• Code Generation Readiness (25%): Machine-readable fields, TASKS-ready metadata

Quality Gate: Score <90% prevents TASKS artifact creation.

5. Threshold Registry Integration

Purpose: Prevent magic numbers by referencing centralized threshold registry.

When @threshold Tag is Required: Use for ALL quantitative values that are:

• Performance configurations (latencies, throughput, IOPS)
• Timeout configurations (connection, read, write timeouts)
• Rate limiting values (requests per second, burst limits)
• Resource limits (memory, CPU, storage)
• Circuit breaker configurations

@threshold Tag Format in YAML:

# String value format
performance:
  p95_latency_ms: "@threshold: PRD.NN.perf.api.p95_latency"

# Comment format for documentation
timeout:
  request_ms: 5000  # @threshold: PRD.NN.timeout.request.sync

Invalid (hardcoded values):

performance:
  p95_latency_ms: 200
timeout:
  request_ms: 5000
rate_limit:
  requests_per_second: 100

Valid (registry references):

performance:
  p95_latency_ms: "@threshold: PRD.NN.perf.api.p95_latency"
timeout:
  request_ms: "@threshold: PRD.NN.timeout.request.sync"
rate_limit:
  requests_per_second: "@threshold: PRD.NN.limit.api.requests_per_second"

6. cumulative_tags Field (CRITICAL)

CRITICAL: SPEC must include cumulative_tags section with ALL upstream tags

Element Type Codes for Cumulative Tags:

Format (maximum - IMPL and CTR included):

cumulative_tags:
  brd: ["BRD.01.01.03", "BRD.01.01.05"]
  prd: ["PRD.01.07.02", "PRD.01.07.15"]
  ears: ["EARS.01.25.01", "EARS.01.25.02"]
  bdd: ["BDD.01.14.01"]
  adr: ["ADR-033", "ADR-045"]
  sys: ["SYS.01.26.01", "SYS.01.26.07"]
  req: ["REQ.01.27.01"]
  impl: ["IMPL.01.29.01"]
  contracts: ["CTR-01"]

Format (minimum - IMPL and CTR skipped):

cumulative_tags:
  brd: ["BRD.01.01.03"]
  prd: ["PRD.01.07.02"]
  ears: ["EARS.01.25.01"]
  bdd: ["BDD.01.14.01"]
  adr: ["ADR-033", "ADR-045"]
  sys: ["SYS.01.26.01"]
  req: ["REQ.01.27.01"]

Tag Count: 7-9 tags (minimum 7, maximum 9)

7. contract_ref Field

Purpose: Link SPEC to CTR (if Layer 9 created)

Format:

interfaces:
  api_endpoints:
    - endpoint: "/api/v1/trades/validate"
      method: "POST"
      contract_ref: "CTR-01"  # Link to contract
      contract_path: "#/paths/~1api~1v1~1trades~1validate/post"  # JSON Pointer

  data_models:
    - model: "TradeOrderRequest"
      schema_ref: "CTR-01#/components/schemas/TradeOrderRequest"

8. Implementation Readiness

100% Implementation-Ready: SPEC must contain ALL information needed to write code

Checklist:

• All modules identified with file paths
• All functions identified with signatures
• All algorithms documented step-by-step
• All data models linked to schemas
• All error codes defined
• All configuration specified
• All tests specified
• Deployment requirements complete

Tag Format Convention (By Design)

The SDD framework uses two distinct notation systems for cross-references:

Key Distinction:

• @adr: ADR-033 → Points to the document ADR-033_risk_limit_enforcement.md
• @brd: BRD.17.01.01 → Points to element 01.01 inside document BRD-017.md

Unified Element ID Format (MANDATORY)

For hierarchical requirements (BRD, PRD, EARS, BDD, SYS, REQ, IMPL):

• Always use: TYPE.NN.TT.SS (dot separator, 4-segment unified format)
• Never use: TYPE-NN:NNN (colon separator - DEPRECATED)
• Never use: TYPE.NN.TT (3-segment format - DEPRECATED)

Examples:

• @brd: BRD.17.01.01 ✅
• @brd: BRD.017.001 ❌ (old 3-segment format)

Validation Checks

Tier 1: Errors (Blocking)

Tier 2: Warnings (Recommended)

Tier 3: Info

Upstream/Downstream Artifacts

Upstream Sources:

• BRD (Layer 1) - Business requirements
• PRD (Layer 2) - Product features
• EARS (Layer 3) - Formal requirements
• BDD (Layer 4) - Test scenarios
• ADR (Layer 5) - Architecture decisions
• SYS (Layer 6) - System requirements
• REQ (Layer 7) - Atomic requirements (PRIMARY SOURCE)
• IMPL (Layer 8) - Implementation approach (optional)
• CTR (Layer 9) - Data contracts (optional)

Downstream Artifacts:

• TASKS (Layer 11) - Task breakdown
• IPLAN (Layer 12) - Implementation plans
• Code (Layer 13) - Implementation

Same-Type Document Relationships (conditional):

• @related-spec: SPEC-NN - SPECs sharing implementation context
• @depends-spec: SPEC-NN - SPEC that must be implemented first

Creation Process

Step 1: Read Upstream Artifacts

Focus on REQ (Layer 7) and optionally CTR (Layer 9).

Step 2: Reserve ID Number

Check docs/SPEC/ for next available ID number (or create docs/SPEC/ directory if first SPEC).

ID Numbering Convention: Start with 2 digits and expand only as needed.

• ✅ Correct: SPEC-01, SPEC-99, SPEC-102
• ❌ Incorrect: SPEC-001, SPEC-009 (extra leading zero not required)

Step 3: Create SPEC File

File naming: docs/SPEC/SPEC-NN_{slug}.yaml

Example: docs/SPEC/SPEC-01_trade_validation.yaml

IMPORTANT: Pure YAML format (NOT markdown)

Note: Templates and examples are in ai_dev_flow/SPEC/ while project-specific SPECs go in docs/SPEC/.

Step 4: Fill Metadata Section

Complete spec_id, title, version, dates, status, owner, task_ready_score.

Step 5: Add Cumulative Tags

Include all 7-9 upstream tags (brd through req/impl/contracts).

Step 6: Define Overview

Purpose, scope, and requirements list.

Step 7: Specify Architecture

Pattern, layers, and technologies (reference ADR decisions).

Step 8: Define Interfaces

API endpoints (with contract_ref), data models (with schema_ref).

Step 9: Document Implementation

Modules (file paths), functions (signatures), algorithms (step-by-step).

Step 10: Specify Error Handling

Error codes, HTTP status, messages, recovery procedures.

Step 11: Define Configuration

Environment variables, feature flags, defaults. Use @threshold for quantitative values.

Step 12: Specify Testing

Unit tests, integration tests, performance tests with targets.

Step 13: Define Deployment

Container, resources, scaling, environment.

Step 14: Add Monitoring

Metrics, alerts, logging requirements.

Step 15: Add Traceability

Upstream sources and downstream artifacts.

Step 16: Create/Update Traceability Matrix

MANDATORY: Create or update docs/SPEC/SPEC-00_TRACEABILITY_MATRIX.md (use template from ai_dev_flow/SPEC/SPEC-00_TRACEABILITY_MATRIX-TEMPLATE.md)

Step 17: Validate SPEC

# YAML validation
yamllint docs/SPEC/SPEC-01_*.yaml

# Schema validation (use SPEC_SCHEMA.yaml as reference)
# Manual: Compare against ai_dev_flow/SPEC/SPEC_SCHEMA.yaml structure

# Cumulative tagging validation
python ai_dev_flow/scripts/validate_tags_against_docs.py --artifact SPEC-01 --expected-layers brd,prd,ears,bdd,adr,sys,req,impl,contracts --strict

# Cross-document validation
python ai_dev_flow/scripts/validate_cross_document.py --document docs/SPEC/SPEC-01_*.yaml

Step 18: Commit Changes

Commit SPEC file and traceability matrix.

Validation

Automated Validation

# YAML validation
yamllint docs/SPEC/SPEC-01_*.yaml

# Schema validation (use SPEC_SCHEMA.yaml as reference)
# Manual: Compare against ai_dev_flow/SPEC/SPEC_SCHEMA.yaml structure

# Cumulative tagging validation
python ai_dev_flow/scripts/validate_tags_against_docs.py \
  --artifact SPEC-01 \
  --expected-layers brd,prd,ears,bdd,adr,sys,req,impl,contracts \
  --strict

# Cross-document validation
python ai_dev_flow/scripts/validate_cross_document.py --document docs/SPEC/SPEC-01_*.yaml

Manual Checklist

• Pure YAML format (not markdown)
• Metadata section complete with task_ready_score
• cumulative_tags section with 7-9 upstream tags
• Overview defines purpose and scope
• Architecture references ADR decisions
• Interfaces link to CTR (if Layer 9 created)
• Implementation specifies modules with file paths
• Functions have signatures and algorithms
• Error handling complete
• Configuration uses @threshold for quantitative values
• Testing requirements defined
• Deployment requirements complete
• Monitoring specified
• Traceability links to upstream/downstream
• 100% implementation-ready
• Element IDs use SPEC.NN.TT.SS format

Diagram Standards

All diagrams MUST use Mermaid syntax. Text-based diagrams (ASCII art, box drawings) are prohibited. See: ai_dev_flow/DIAGRAM_STANDARDS.md and mermaid-gen skill.

Common Pitfalls

1. Markdown format: SPEC must be pure YAML, not markdown
2. Missing cumulative_tags: Must include all 7-9 upstream tags
3. No contract_ref: Must link to CTR if Layer 9 created
4. Vague implementation: Must specify exact file paths and signatures
5. Missing algorithms: Functions need step-by-step algorithms
6. Incomplete: Must be 100% implementation-ready
7. Hardcoded values: Use @threshold for performance/timeout/rate limits
8. Wrong element IDs: Use SPEC.NN.TT.SS, not legacy STEP-XXX, IF-XXX, DM-XXX
9. Wrong cumulative tag codes: Use correct element type codes (EARS=25, BDD=14, SYS=26, REQ=27, IMPL=29)

Post-Creation Validation (MANDATORY - NO CONFIRMATION)

CRITICAL: Execute this validation loop IMMEDIATELY after document creation. Do NOT proceed to next document until validation passes.

Automatic Validation Loop

LOOP:
  1. Run: python ai_dev_flow/scripts/validate_cross_document.py --document {doc_path} --auto-fix
  2. IF errors fixed: GOTO LOOP (re-validate)
  3. IF warnings fixed: GOTO LOOP (re-validate)
  4. IF unfixable issues: Log for manual review, continue
  5. IF clean: Mark VALIDATED, proceed

Validation Command

# Per-document validation (Phase 1)
python ai_dev_flow/scripts/validate_cross_document.py --document docs/SPEC/SPEC-NN_slug.yaml --auto-fix

# Layer validation (Phase 2) - run when all SPEC documents complete
python ai_dev_flow/scripts/validate_cross_document.py --layer SPEC --auto-fix

Layer-Specific Upstream Requirements

Auto-Fix Actions (No Confirmation Required)

Validation Codes Reference

Quality Gate

Blocking: YES - Cannot proceed to next document until Phase 1 validation passes with 0 errors.

Next Skill

After creating SPEC, use:

doc-tasks - Create Task Breakdown (Layer 11)

The TASKS will:

• Reference this SPEC as upstream source
• Include all 8-10 upstream tags
• Break SPEC into actionable tasks
• Provide AI-structured TODO format

Reference Documents

SPEC artifacts do not support REF documents. Reference documents are limited to BRD and ADR types only per the SDD framework.

For supplementary documentation needs, create:

• BRD-REF: Business context documentation
• ADR-REF: Technical reference guides (API quick references, implementation guides)

Related Resources

• Template: ai_dev_flow/10_SPEC/SPEC-TEMPLATE.yaml (primary authority)
• SPEC Creation Rules: ai_dev_flow/10_SPEC/SPEC_CREATION_RULES.md
• SPEC Validation Rules: ai_dev_flow/10_SPEC/SPEC_VALIDATION_RULES.md
• SPEC Schema: ai_dev_flow/10_SPEC/SPEC_SCHEMA.yaml
• SPEC README: ai_dev_flow/10_SPEC/README.md
• Shared Standards: .claude/skills/doc-flow/SHARED_CONTENT.md

Quick Reference

SPEC Purpose: Implementation-ready technical specifications

Layer: 10

Element ID Format: SPEC.NN.TT.SS

• Step = 15
• Interface = 16
• Data Model = 17
• Validation Rule = 21
• Specification Element = 28

Removed Patterns: STEP-XXX, IF-XXX, INT-XXX, DM-XXX, MODEL-XXX, VR-XXX

Tags Required: @brd through @req/impl/contracts (7-9 tags)

Format: Pure YAML (not markdown)

Key Features:

• cumulative_tags section (CRITICAL)
• contract_ref field (links to CTR)
• schema_ref field (links to data models)
• @threshold references for quantitative values
• 100% implementation-ready
• All modules, functions, algorithms specified

TASKS-Ready Score: ≥90% required for "approved" status

Quality Gate: Must be 100% implementation-ready

Next: doc-tasks