Skip to main content

document-a-software-architecture

Document a software system's boundaries, components, data, decisions, runtime, security, failure modes, and change paths. Use when maintainers or reviewers need a current architecture guide grounded in the real implementation.

Source facts

Repository
protosphinx/sphinxstack
Last source activity
July 29, 2026 at 13:55
Detected SKILL.md language
English
Stars
3
Forks
0

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer
2 files

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
name
document-a-software-architecture
category
write
description
Document a software system's boundaries, components, data, decisions, runtime, security, failure modes, and change paths. Use when maintainers or reviewers need a current architecture guide grounded in the real implementation.
# document-a-software-architecture Explain the system at the level needed to change and operate it. Prefer a small set of linked views and decision records over one diagram that claims to show everything. ## Inputs - Gather repositories, deployment configuration, schemas, interfaces, infrastructure, observability, and ownership. - Review current runtime evidence, dependency inventories, incidents, threat models, and architecture decisions. - Define the audience and the maintenance or review questions the document must answer. ## Procedure 1. State system purpose, users, scope, non-goals, quality attributes, and important constraints. 2. Draw context and trust boundaries around people, external systems, and owned services. 3. Map major components, responsibilities, interfaces, dependencies, and owners. 4. Trace critical request, event, data, and control flows end to end. 5. Document persistent data, lifecycle, consistency, privacy, retention, and recovery. 6. Explain deployment units, environments, configuration, secrets, scaling, and network boundaries. 7. Record availability, security, capacity, observability, and operational assumptions. 8. Describe expected failures, containment, degradation, retry, reconciliation, and recovery behavior. 9. Link consequential decisions and known debt rather than rewriting their full history. 10. Verify diagrams and claims against code, deployed configuration, and responsible owners; add freshness triggers. ## Boundaries Do not publish secrets, private endpoints, customer data paths, or exploitable detail beyond the document's approved audience. Never document intended architecture as current without labeling the gap. ## Done - Context, components, interfaces, data, deployment, trust, and ownership views are linked - Critical flows and failure behavior are checked against implementation and runtime evidence - Constraints, decisions, debt, and intended changes are visibly distinguished - The document has an owner and review triggers tied to system change
View on GitHub