AGENT SKILL

well-architected

Review an architecture, change, or IaC against the six AWS Well-Architected pillars — Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability — and produce evidence-backed, severity-tagged findings. Use for architecture reviews, design feedback, or when the user mentions "Well-Architected" / "WA review" or asks about best practices for reliability, security, cost, performance, or sustainability.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 0
Forks
0

01 前言 Front Matter

name
well-architected
description
Review an architecture, change, or IaC against the six AWS Well-Architected pillars — Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability — and produce evidence-backed, severity-tagged findings. Use for architecture reviews, design feedback, or when the user mentions "Well-Architected" / "WA review" or asks about best practices for reliability, security, cost, performance, or sustainability.

03 SKILL.md 正文

Well-Architected Review

Apply the AWS Well-Architected Framework when reviewing architectures, writing code, or advising on design decisions. Evaluate every pillar systematically, cite evidence from the code/IaC, and rank findings by risk.

Adapted from the AWS sample-well-architected-skills-and-steering steering document (github.com/aws-samples/sample-well-architected-skills-and-steering), condensed into a self-contained, provider-agnostic rubric. The pillars and principles are general cloud-design guidance, not tied to any one platform.

When to Apply

Apply this skill whenever the user:

  • Asks for an architecture review or design feedback
  • Requests help designing a new workload or system
  • Asks about best practices for reliability, security, cost, performance, or sustainability
  • Mentions "Well-Architected" or "WA review"

Pillars

Always consider all six pillars when evaluating or proposing architectures:

  1. Operational Excellence — Automate operations, make frequent small reversible changes, refine procedures, anticipate failure, learn from operational events. (Observability, deployment risk, operational readiness, event/incident management.)
  2. Security — Implement a strong identity foundation, enable traceability, apply security at all layers, automate security best practices, protect data in transit and at rest, keep people away from data, prepare for security events. (Identity, permissions, detection, network/compute protection, data protection, incident response, application security.)
  3. Reliability — Automatically recover from failure, test recovery procedures, scale horizontally, stop guessing capacity, manage change through automation. (Quotas, network topology, distributed-system design, monitoring, scaling, backups, fault isolation, DR.)
  4. Performance Efficiency — Democratize advanced technologies, go global in minutes, use serverless architectures, experiment more often, consider mechanical sympathy. (Resource selection, compute, data/storage, networking, caching, the optimization process.)
  5. Cost Optimization — Implement cloud financial management, adopt a consumption model, measure overall efficiency, stop spending on undifferentiated heavy lifting, analyze and attribute expenditure. (Usage governance, monitoring, decommissioning, right-sizing, pricing models, data transfer, demand management.)
  6. Sustainability — Understand your impact, establish sustainability goals, maximize utilization, anticipate and adopt more efficient offerings, use managed services, reduce downstream impact. (Region selection, demand alignment, efficient architecture patterns, data management, hardware selection.)

Design Principles

When proposing solutions:

  • Favor managed services over self-managed infrastructure
  • Design for failure — assume any component can fail at any time
  • Decouple components to reduce blast radius
  • Use multiple Availability Zones (or equivalent fault domains) for high availability
  • Implement least-privilege access for all identities
  • Automate everything that can be automated
  • Use infrastructure as code for all environments
  • Design for observability from day one

Review Approach

  1. Identify the workload scope and business context.
  2. Evaluate each pillar systematically — do not skip pillars. If a pillar cannot be assessed from the evidence available, mark it N/A with a one-line reason rather than inventing findings.
  3. Identify high-risk issues (HRIs) and medium-risk issues (MRIs).
  4. Prioritize findings by business impact and effort to remediate.
  5. Provide specific, actionable recommendations, citing the file/line and a concrete service or pattern to adopt — never "improve security."
  6. Reference a relevant Well-Architected lens if the workload fits one (Serverless, SaaS, Data Analytics, Machine Learning, IoT, Generative AI, etc.).

Trade-off Guidance

Acknowledge trade-offs explicitly:

  • Security controls may add latency — quantify the impact
  • High availability increases cost — present options at different tiers
  • Performance optimization may reduce portability — state the lock-in risk
  • Cost optimization may reduce resilience — make the risk visible

Response Format

When delivering guidance:

  • Lead with the most critical finding or recommendation
  • Group findings by pillar
  • Use severity labels: 🔴 High Risk, 🟡 Medium Risk, 🟢 Best Practice
  • Include "Why it matters" for each finding
  • Provide a concrete next step for each recommendation

04 同仓库更多 Skills

fabiocicerchia
fabiocicerchia/gandalf

security-assessment

Use to generate a CNCF TAG Security style self-assessment for a software project. Produces a structured document covering metadata, overview, actors, actions, goals, compliance, secure development practices, and incident response.

★ 0
fabiocicerchia
fabiocicerchia/gandalf

ruthless-refactor

Aggressively simplify and shrink an inflated codebase without changing behavior — find and delete dead code, collapse duplication, remove needless indirection and over-engineering, and replace custom code with language or library features. Use when the user says things like "simplify this module", "this codebase is bloated", "reduce the line count", "remove dead code", "there's too much abstraction here", or asks to make code smaller and leaner while keeping it fully working. Works in small test-verified passes and reports what was removed and why it was safe. Not for bug-hunting or scored review — use a code-review or quality-gate skill for that.

★ 0
fabiocicerchia
fabiocicerchia/gandalf

quality-gate-review

Review a code change, PR, diff, or code-challenge submission against fixed quality gates and produce a scored go/no-go decision. Runs six gates (correctness & algorithm, tests & CI, security & robustness, clean code & design, readability & documentation, project & delivery hygiene) built from an 82-item review checklist, scores each 0-5, computes a weighted overall score out of 100, gives a GO / REVIEW / NO-GO verdict, and lists the specific blockers to fix. Use whenever the user wants to judge whether work is ready to move forward — phrases like "review my work", "score this PR", "review this code challenge", "is this ready to ship/merge", "run the quality gates", "should I move forward", "rate the quality", "grade this submission". This is a scored decision aid, not a line-by-line bug hunt; for pure bug-finding use a code-review tool instead, and lean on this when the user wants a quality level and a ship-or-not call.

★ 0
fabiocicerchia
fabiocicerchia/gandalf

pr-code-summarizer

Use to summarise a pull request or diff for a technical leader. Produces a 60-second readable summary with risks and questions.

★ 0
fabiocicerchia
fabiocicerchia/gandalf

improve-codebase-architecture

Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.

★ 0
fabiocicerchia
fabiocicerchia/gandalf

grilling

Grill the user relentlessly about a plan or design. Use when the user wants to stress-test a plan before building, or uses any 'grill' trigger phrases.

★ 0