Governance Standards Principles Coding

Coding

Code quality standards for generated application and infrastructure code.

Domain: principles

---

Checks (5)

Check	Description
STAN-CODE-001	Meaningful Names: Use descriptive, intention-revealing names for variables, functions, resources, and modules. Avoid abbreviations unless they are universally understood (e.g., HTTP, URL, ID).
STAN-CODE-002	Small Functions: Functions should be short and focused. If a function exceeds roughly 30 lines, it likely does too much and should be split.
STAN-CODE-003	Error Handling at Boundaries: Handle errors at system boundaries (user input, external APIs, file I/O). Internal code should trust its contracts.
STAN-CODE-004	Consistent Module Structure: Terraform modules should follow a consistent file layout: main.tf, variables.tf, outputs.tf, locals.tf, providers.tf. Bicep modules should separate parameters, resources, and outputs.
STAN-CODE-005	Parameterize, Don't Hard-Code: All environment-specific values (names, regions, SKUs, IP ranges) must be parameterized via variables or parameters. Never embed environment-specific values in resource definitions.

---

STAN-CODE-001

Meaningful Names: Use descriptive, intention-revealing names for variables, functions, resources, and modules. Avoid abbreviations unless they are universally understood (e.g., HTTP, URL, ID).

Rationale: Parameterized configurations allow reuse across environments and prevent hardcoded values.
Agents: terraform-agent, bicep-agent, app-developer, csharp-developer, python-developer, react-developer

Examples

• Use 'storage_account' not 'sa'; 'container_registry' not 'cr'
• Use 'get_user_by_email()' not 'get_u()'

---

STAN-CODE-002

Small Functions: Functions should be short and focused. If a function exceeds roughly 30 lines, it likely does too much and should be split.

Rationale: The isolated worker model provides better dependency isolation and .NET version flexibility.
Agents: app-developer, csharp-developer, python-developer, react-developer

Examples

• Split 'process_order()' into 'validate_order()', 'calculate_total()', 'save_order()'

---

STAN-CODE-003

Error Handling at Boundaries: Handle errors at system boundaries (user input, external APIs, file I/O). Internal code should trust its contracts.

Rationale: Parameterized configurations allow reuse across environments and prevent hardcoded values.
Agents: app-developer, csharp-developer, python-developer, react-developer

Examples

• Validate API request payloads at the controller layer, not in every function
• Wrap external HTTP calls in try/except, not internal method calls

---

STAN-CODE-004

Consistent Module Structure: Terraform modules should follow a consistent file layout: main.tf, variables.tf, outputs.tf, locals.tf, providers.tf. Bicep modules should separate parameters, resources, and outputs.

Rationale: Consistent file organization makes code reviewable and prevents merge conflicts across stages.
Agents: terraform-agent, bicep-agent

Examples

• Terraform: variables.tf for inputs, main.tf for resources, outputs.tf for outputs
• Bicep: param block at top, resource declarations, output block at bottom

---

STAN-CODE-005

Parameterize, Don't Hard-Code: All environment-specific values (names, regions, SKUs, IP ranges) must be parameterized via variables or parameters. Never embed environment-specific values in resource definitions.

Rationale: Parameterized configurations allow reuse across environments and prevent hardcoded values.
Agents: terraform-agent, bicep-agent, app-developer, csharp-developer, python-developer, react-developer

Examples

• Terraform: use var.location not 'eastus' in resource blocks
• Bicep: use param location string not 'eastus'
• Application: use environment variables or config files, not inline strings

---