Guide AI coding agents to make code changes that preserve architectural integrity, minimize complexity, and are verified with evidence.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install engineering-architecture-memory
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install engineering-architecture-memory using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Engineering & Architecture Memory?
Engineering & Architecture Memory is a persistent set of operating principles for AI coding agents working on codebases, feature designs, and code reviews. It prompts the agent to understand architecture, boundaries, conventions, quality attributes, and trade-offs before acting—helping Openclaw Skills users get changes that fit the system rather than merely satisfy a narrow request.
The skill emphasizes simple designs, backward compatibility, security, and measurable verification. It also asks agents to distinguish concrete engineering risks from personal preferences, review their own changes critically, and avoid unnecessary rewrites or scope expansion.
Engineering & Architecture Memory Use Cases
- Before modifying an existing codebase, to inspect its architecture, patterns, dependencies, and boundaries.
- When designing a feature, to translate outcomes, constraints, and acceptance criteria into the simplest viable design.
- During code review, to identify security risks, coupling, boundary violations, unnecessary complexity, and introduced technical debt.
- When a change could affect reliability, performance, scalability, maintainability, testability, or other quality attributes.
- Before committing to an uncertain architectural decision, to validate assumptions with a focused test, prototype, or benchmark.
- When preserving existing APIs and behavior or controlling the blast radius of a change is important.
How Engineering & Architecture Memory Works
- Understand the task in context: inspect the existing architecture, dependencies, conventions, constraints, and quality requirements.
- Audit relevant code before changing it; identify established abstractions and distinguish concrete issues from stylistic preferences.
- Define the desired outcome and acceptance criteria, including relevant non-functional requirements such as security, performance, or reliability.
- Choose the smallest design that satisfies the requirements, respecting component boundaries and preferring existing infrastructure over speculative abstractions.
- Implement the change while preserving valid behavior and APIs unless the task requires otherwise; avoid unrelated refactors and unnecessary migrations.
- Review the result for security issues, architectural flaws, inconsistent patterns, excess complexity, and unintended side effects.
- Verify important properties with appropriate evidence, such as tests, static analysis, type checks, dependency checks, benchmarks, or security checks.
- Complete the task with relevant documentation when decisions are non-obvious, and make trade-offs explicit where multiple designs are viable.
Engineering & Architecture Memory Setup
This skill is a guidance document, not a CLI package. Its frontmatter identifies it as engineering-architecture at version 1.0.0; the source does not specify an installation command, dependencies, or additional configuration. Make the document available to the coding agent as a skill and load it as persistent operating guidance for code changes, feature design, and reviews.
# No installation command is specified in the supplied SKILL.md.
# Add the skill to your agent's supported skills location using its normal workflow.
Engineering & Architecture Memory Data Schema & Taxonomy
The skill defines guidance and metadata; it does not prescribe application data formats or generate project files.
| Element | Details |
|---|---|
| Skill name | engineering-architecture |
| Version | 1.0.0 |
| Description | Apply persistent engineering and architectural principles to code changes, feature design, and code review. |
| Main content | Fourteen operating principles plus a core principle focused on correctness, simplicity, architectural integrity, and measurable evidence. |
| Generated files or data | None specified. |
| Metadata taxonomy | YAML frontmatter with name, description, and version; the remaining content is Markdown guidance. |
Engineering & Architecture Memory Advanced Features
- Architecture-first reasoning: Examine dependencies, boundaries, patterns, and quality attributes before proposing implementation changes.
- Evidence-based verification: Encourage tests, static analysis, type checks, dependency checks, performance measurements, security checks, and architecture-specific assertions where practical.
- Security-aware review: Consider trust boundaries, authorization, validation, data exposure, secrets, injection, SSRF, external integrations, and privilege escalation.
- Controlled change scope: Preserve existing behavior and APIs where required, avoid unrelated refactors, and limit migration and compatibility risks.
- Trade-off analysis: Evaluate complexity, operational cost, performance, reliability, maintainability, scalability, and developer experience.
- Uncertainty reduction: Use the smallest useful experiment, prototype, benchmark, or focused test before committing to a larger architectural decision.
- Reviewer mindset: Inspect agent-produced changes for flaws and technical debt without expanding the task unnecessarily.
- Persistent guidance: Apply these principles across relevant tasks unless an explicit task instruction overrides them, making the skill reusable within Openclaw Skills workflows.
SKILL.md
Loading
Related Openclaw Skills
Obsidian Documentation Workflow
A structured Obsidian vault workflow for creating, classifying, validating, linking, updating, and archiving engineering documentation.
StratPit
StratPit lets AI agents play one-on-one strategy games for USDC prizes, with code-enforced rules and automated payouts.
Douyin Competitor Radar
Analyze public Douyin content, creator activity, hot trends, and comments to uncover competitor insights and practical marketing opportunities.
Kuaishou Competitor Radar
Collect and compare public Kuaishou video, creator, and comment data to support competitor research and content strategy.
Sandbox Implementation SOP (SOP-DEV-001)
A mandatory agent workflow for implementing and validating changes in a clean sandbox before anything is executed on the host.
Sandbox Cleanup After Execution
A controlled cleanup procedure that resets a development sandbox after a task while preserving approved artifacts and recording readiness for the next run.
Featured*
