wellarchitected-review-operator
Operates AWS Well-Architected Tool reviews end-to-end — creates workloads, runs pillar reviews (operational excellence, security, reliability, performance, cost optimization, sustainability, Prosperity), answers questions with risk-tier rationale, generates reports and improvement plans, creates milestones, and integrates with Trusted Advisor and Well-Architected Labs. Runs deterministic pre-checks (workload existence, lens availability, pillar coverage), emits the exact update-answer / get-consolidated-report CLI behind a CONFIRM gate, and verifies risk-tier transitions. Emits READY | BLOCKED | COMPLETED. Use when creating a workload, answering pillar questions, generating an improvement plan, creating a milestone, or integrating with Trusted Advisor. Triggers: well-architected tool, well architected review, workload review, pillar, operational excellence, security, reliability, performance, cost optimization, sustainability, prosperity, improvement plan, milestone, trusted advisor.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-24
- Stars
- ★ 4
- Forks
- 0
01 前言 Front Matter
- name
- wellarchitected-review-operator
- license
- Apache-2.0
- metadata
- { "tags": "aws, well-architected, governance, operate, review, pillar, improvement-plan", "phase": "3", "author": "Jacky Chan — AWS Community Builder", "domain": "aws-cloudops", "family": "Governance", "version": "0.1.0", "keywords": "aws, well-architected, wellarchitected, governance, cloudops, operate, review, pillar, operational-excellence, security, reliability, performance, cost-optimization, sustainability, prosperity, improvement-plan, milestone, trusted-advisor, lens", "task_type": "operate", "complexity": "high", "entry_point": "false", "skill_class": "capability", "when_to_use": "Invoke when the user wants to create or operate a Well- Architected Tool workload (name, environment, description, lenses), answer pillar questions (operational excellence, security, reliability, performance efficiency, cost optimization, sustainability, Prosperity), generate a consolidated report, draft an improvement plan, create a milestone, integrate with Trusted Advisor, or pull Well-Architected Labs runbooks. Do NOT invoke for plain AWS Config conformance packs (use config-rule-deployer), for Security Hub control finding triage (use securityhub-control-compliance-auditor), or for audit evidence collection (use auditmanager-assessment-auditor).", "dependencies": "aws-orchestrator", "requires_llm": "true", "verdict_shape": "READY | BLOCKED | COMPLETED", "lifecycle_status": "active", "supports_pipeline": "true" }
- description
- Operates AWS Well-Architected Tool reviews end-to-end — creates workloads, runs pillar reviews (operational excellence, security, reliability, performance, cost optimization, sustainability, Prosperity), answers questions with risk-tier rationale, generates reports and improvement plans, creates milestones, and integrates with Trusted Advisor and Well-Architected Labs. Runs deterministic pre-checks (workload existence, lens availability, pillar coverage), emits the exact update-answer / get-consolidated-report CLI behind a CONFIRM gate, and verifies risk-tier transitions. Emits READY | BLOCKED | COMPLETED. Use when creating a workload, answering pillar questions, generating an improvement plan, creating a milestone, or integrating with Trusted Advisor. Triggers: well-architected tool, well architected review, workload review, pillar, operational excellence, security, reliability, performance, cost optimization, sustainability, prosperity, improvement plan, milestone, trusted advisor.
- compatibility
- Agent runtime that reads SKILL.md (Claude Code, Cursor, Windsurf, Codex, Gemini). No AWS CLI required for offline plan classification. Live-account operations use aws wellarchitected create-workload, update-workload, get-workload, list-workloads, get-answer, update-answer, get-consolidated-report, create-milestone, list-milestones, associate-logs, import-lens, list-lenses, and aws trustedadvisor (AWS CLI v2, SSO or key-based credentials).
02 资源文件 19
- 参考文档eval/test-cases.yaml5.2 KB
- 参考文档evals/baselines/war-answer-security-pillar-high-risk.md528 B
- 参考文档evals/baselines/war-blocked-lens-not-imported.md411 B
- 参考文档evals/baselines/war-create-workload-all-pillars.md576 B
- 参考文档evals/baselines/war-milestone-and-consolidated-report.md558 B
- 参考文档evals/baselines/war-trusted-advisor-integration.md520 B
- 参考文档evals/eval_config.yaml1.8 KB
- 参考文档evals/evals.json4.3 KB
- 参考文档evals/prompts/war-answer-security-pillar-high-risk.md992 B
- 参考文档evals/prompts/war-blocked-lens-not-imported.md959 B
- 参考文档evals/prompts/war-create-workload-all-pillars.md1.2 KB
- 参考文档evals/prompts/war-milestone-and-consolidated-report.md944 B
- 参考文档evals/prompts/war-trusted-advisor-integration.md1.0 KB
- 参考文档example/README.md7.4 KB
- 参考文档references/advanced-patterns.md8.2 KB
- 参考文档references/diagnostic-commands.md4.8 KB
- 参考文档references/pillar-question-bank.md13.5 KB
- 参考文档references/review-cli-commands.md11.2 KB
- 参考文档skill.usf.yaml4.3 KB
03 SKILL.md 正文
Well-Architected Review Operator
An AWS CloudOps agent skill that operates Well-Architected Tool reviews safely — creates workloads, runs pillar reviews, answers questions with risk-tier rationale, generates consolidated reports and improvement plans, creates milestones, and integrates with Trusted Advisor and Well-Architected Labs. Runs deterministic pre-checks and emits READY | BLOCKED | COMPLETED with the exact CLI sequence and post-verification.
Quick navigation
| Need | Section |
|---|---|
| What MUST be in the response | "STRICT output contract" |
| Why the review order matters | "Reasoning framework" |
| Pre-flight checks before any CLI | "Pre-flight" |
| Pillar-by-pillar procedure | "Review procedure" |
| Common silent-failure pitfalls | "NEVER" |
| Risk-tier answer strategy | "Expert heuristic" |
| Prosperity pillar (2025) | "Recent AWS features" |
| Trusted Advisor integration | "Trusted Advisor integration" |
| Deep CLI sequences | references/review-cli-commands.md |
| Pillar question bank + improvement plan patterns | references/pillar-question-bank.md |
STRICT output contract
When this skill is invoked with a Well-Architected review operation (create-workload, answer-pillar, generate-report, create-milestone, integrate-trusted-advisor, or a partial existing workload state), the agent MUST respond with the WORKLOAD / VERDICT / TARGET / PRE_CHECKS / CHECKLIST / RISK_COUNTS / IMPROVEMENT_PLAN / STEPS / POST_VERIFY / STATE / NOTES block defined in "Output format" using the literal all-caps labels. Do NOT preface the block with prose, headings, or disclaimers — emit the block as the first lines of the response.
Required output structure
WORKLOAD: <workload-name>— the workload display name.VERDICT: READY | BLOCKED | COMPLETEDTARGET: <workload-id, lens, pillar, region>PRE_CHECKS:followed by indented[PASS]/[FAIL]lines.CHECKLIST:followed by onePILLAR:row per pillar (Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability), each with risk counts (HIGH=<n> MEDIUM=<n> NONE=<n>) and per-question risk-tier lines citing workload evidence.RISK_COUNTS:with aggregateHIGH:,MEDIUM:,NONE:totals across all pillars.IMPROVEMENT_PLAN:followed by numbered items, each tagged[HIGH]or[MEDIUM], with gap description, owner, target date, and remediation link.STEPS:followed by indented commands (CONFIRM gate first for state-changing operations).POST_VERIFY:followed by indented[PASS]/[FAIL]lines.STATE:/NOTES:with rationale and caveats.
FORBIDDEN output patterns
- NEVER preface the block with prose —
WORKLOAD:is the FIRST line, always. No greetings, no "Let me analyze…", no disclaimers before the block. Use uppercase verdict values only (READY,BLOCKED,COMPLETED). - NEVER use markdown variants of labels — write
VERDICT:, not**VERDICT:**,### Verdict, or`VERDICT`. - NEVER swap verdict tokens — exactly
READY,BLOCKED,COMPLETED. Not "ok", "done", "pending", "ERROR". - NEVER omit a pillar from CHECKLIST — all six pillars (Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability) must appear, even if risk counts are zero.
- NEVER list a risk without citing workload evidence —
every
HIGH_ISSUEandMEDIUM_ISSUEline must include the specific workload gap that justifies the tier. - NEVER omit PRE_CHECKS — every pre-check must appear with
[PASS]or[FAIL]and a specific reason for each failure. - NEVER emit placeholder values in a COMPLETED plan — every flag, ARN, and question ID populated with actual values from the input data.
- **NEVER claim COMPLETED without every POST_VERIFY line showing
[PASS], and never omit the CONFIRM gate as the first STEPS entry for state-changing operations.
Reasoning framework (why the review order matters)
Why the review order matters (lens binding, environment risk tolerance, milestone snapshots, TA as evidence): Advanced patterns. Load on demand when sequencing an unfamiliar review.
Pre-flight: workload / lens metadata gate
Pre-flight metadata gate commands (list-workloads / list-lenses / get-workload / list-answers / list-share-invitations / describe-checks / sts get-caller-identity): Diagnostic commands. Load on demand before classification in live-account runs.
Review procedure (apply in order)
Step 1: Create or confirm the workload
aws wellarchitected create-workload \
--workload-name "checkout-service" \
--description "Checkout microservice handling payment \
authorization and order capture" \
--environment PRODUCTION \
--review-owner "payments-platform@example.com" \
--lenses wellarchitected \
--aws-regions us-east-1 us-west-2 \
--account-ids 111122223333 \
--pillar-ids "performance" "security" "reliability" \
"operationalExcellence" "costOptimization" "sustainability" \
--region us-east-1
Capture WorkloadId from the response. Every subsequent
update-answer and get-consolidated-report requires it.
Lens selection — wellarchitected (the default Framework)
covers the six base pillars. To apply specialty lenses (SaaS,
FTR, Healthcare, Prosperity), add the lens alias to --lenses:
--lenses wellarchitected wellarchitected-saas wellarchitected-prosperity
Specialty lenses must be enabled in the account first via
import-lens. The Prosperity lens (wellarchitected-prosperity)
ships the new Prosperity pillar (2025) covering sustainability
of growth, customer-centric measures, and value-stream health.
Step 2: Confirm lens availability
aws wellarchitected list-lenses --workload-id <id> --region us-east-1
aws wellarchitected list-lenses --region us-east-1 # all available lenses
If a lens is missing from the workload but listed in the
account, call associate-lenses:
aws wellarchitected associate-lenses \
--workload-id <id> \
--lens-aliases wellarchitected-prosperity \
--region us-east-1
If a lens is not in the account at all, call import-lens:
aws wellarchitected import-lens \
--lens-alias wellarchitected-prosperity \
--region us-east-1
Step 3: Answer pillar questions (risk-tier rationale)
For each pillar, answer the canonical question set with a
risk-tier rationale. The Well-Architected Tool stores answers
as ChoiceUpdates — a map of choice IDs to risk-tier +
rationale + notes.
aws wellarchitected update-answer \
--workload-id <id> \
--lens-alias wellarchitected \
--question-id <question-id> \
--choice-updates '{
"<choice-id>": {
"Status": "SELECTED",
"Reason": "OUT_OF_SCOPE",
"Notes": "Choice not applicable — workload is serverless with no OS patching"
},
"<risky-choice-id>": {
"Status": "NOT_SELECTED",
"Reason": "RISK_GUIDANCE",
"Notes": "No automated rollback — HIGH risk per APP-SVC rollback SOP"
}
}' \
--region us-east-1
The skill identifies the appropriate choice updates per
question and emits the rationale tied to actual workload
evidence (deployment topology, observability coverage,
on-call runbooks). The pillar question bank in
references/pillar-question-bank.md enumerates the canonical
questions per pillar and the risk-tier mapping.
Risk-tier semantics:
HIGH_ISSUE— significant gap; the workload is at risk in this dimension. Goes to the top of the improvement plan.MEDIUM_ISSUE— moderate gap; should be addressed but not urgent.NO_ISSUE— pillar satisfied; no action needed.OUT_OF_SCOPE— choice not applicable to the workload (e.g., OS patching on a serverless function).
Step 4: Generate the consolidated report
aws wellarchitected get-consolidated-report \
--workload-id <id> \
--format JSON \
--include-shared-resources \
--region us-east-1 > /tmp/war-report-$(date +%s).json
The consolidated report enumerates:
- Per-pillar risk distribution (HIGH / MEDIUM / NONE counts).
- Improvement plan items (one per
HIGH_ISSUE/MEDIUM_ISSUEanswer). - Lens-specific findings (specialty lenses appear here).
- Workload metadata and review owner.
PDF report variant CLI: Diagnostic commands. Load on demand when an executive-ready PDF is requested.
Step 5: Draft the improvement plan
The improvement plan is derived from every HIGH_ISSUE and
MEDIUM_ISSUE answer. The skill surfaces the plan as a
prioritized list with the underlying question, the gap
evidence, and a remediation pointer (often a Well-Architected
Labs runbook):
aws wellarchitected list-improvement-plans \
--workload-id <id> \
--lens-alias wellarchitected \
--region us-east-1
Per-pillar remediation link mapping (operational excellence / security / cost labs): Advanced patterns. Load on demand when augmenting improvement plan items.
Step 6: Create a milestone
aws wellarchitected create-milestone \
--workload-id <id> \
--milestone-name "2026-Q3-baseline" \
--region us-east-1
A milestone snapshots the current answers and improvement plan at the moment of creation. Subsequent answer updates do NOT back-propagate. Create the milestone AFTER the review is complete and the improvement plan is drafted.
list-milestones CLI: Diagnostic commands. Load on demand when verifying milestones.
Step 7: Integrate Trusted Advisor findings
TA pulling CLI and the TA-check → pillar-question mapping: Diagnostic commands. Load on demand when integrating Trusted Advisor evidence.
Step 8: Wire Well-Architected Labs runbooks
Per-pillar labs wiring list (including Prosperity value-stream labs): Advanced patterns. Load on demand when attaching runbooks to plan items.
Step 9: Verify review completeness
Per-pillar completeness verification CLI and the UNANSWERED rule: Diagnostic commands. Load on demand before declaring a review complete.
Step 10: Tag, share, and close
tag-resource and cross-account create-workload-share CLI: Diagnostic commands. Load on demand at review close-out.
Pillar matrix
| Pillar | ID | Focus | Specialty lens |
|---|---|---|---|
| Operational Excellence | operationalExcellence |
Run workloads effectively, monitor, automate | — |
| Security | security |
IAM, detection, infrastructure protection, data protection | AWS Well-Architected Security lens |
| Reliability | reliability |
Recover from failures, meet demand, mitigate disruptions | — |
| Performance Efficiency | performance |
Compute, storage, database, networking, trade-offs | — |
| Cost Optimization | costOptimization |
Resource selection, supply/demand, optimization | — |
| Sustainability | sustainability |
Environmental impact, resource utilization, ROI | — |
| Prosperity (2025) | (via Prosperity lens) | Growth sustainability, customer-centric measures, value streams | wellarchitected-prosperity |
Trusted Advisor integration
TA check category table and findings-pulling CLI: Diagnostic commands. Load on demand when mapping TA evidence to answers.
Edge-case handling
Full edge-case catalog (lens import, environment mismatch, silent answer loss, milestone snapshots, empty reports, share rejection, Prosperity lens, TA migration): Advanced patterns. Load on demand when a review operation fails or silently no-ops.
Recent AWS features (2024-2026)
2024-2026 feature delta (Prosperity lens, API expansion, AWS Health Aware TA feed, Labs IaC, custom lenses, cross-account sharing, PDF reports, Resilience Hub import): Advanced patterns. Load on demand when a question references a recent capability.
NEVER (top 5 — full list in references)
- NEVER answer a question with a risk-tier without citing
workload evidence. A bare
HIGH_ISSUEwith no rationale is rejected by the operator as invalid and pollutes the improvement plan. - NEVER create a milestone before the review is complete. A milestone is a point-in-time snapshot; answer updates after the milestone do NOT back-propagate. Create the milestone AFTER the review and improvement plan are drafted.
- NEVER reference a specialty lens without first running
import-lensANDassociate-lenses. An un-imported lens producesResourceNotFoundExceptionatupdate-answertime. - NEVER treat Trusted Advisor findings as auto-answers. TA is an evidence source; the operator must confirm the risk-tier. Auto-answering with TA findings produces false confidence.
- NEVER omit a pillar from
--pillar-idsat workload creation. A missing pillar produces an incomplete consolidated report and gaps the improvement plan.
Expert heuristic — risk-tier strategy
Full risk-tier strategy (evidence-driven answers, HIGH-risk owners, PREPROD tolerance, TA as evidence, milestone cadence, lens cross-reference): Advanced patterns. Load on demand when deciding risk tiers.
Pre-flight safety checks (run before any state-changing CLI)
- MANDATORY CONFIRMATION GATE before any state-changing
CLI (
create-workload,update-answer,create-milestone,associate-lenses,import-lens,create-workload-share). - Snapshot before modification:
get-workload --workload-id <id> --output json > /tmp/<id>-backup-$(date +%s).json. - Verify lens availability:
list-lenses --workload-id <id>shows the lens is associated. If missing, runimport-lensassociate-lensesfirst.
- Verify the operator holds the ReviewOwner role: the
workload
IsReviewOwnerUpdateAllowedmust betrue. - Verify the lens question ID before update-answer: wrong choice IDs silently fail to persist the answer.
- Verify Trusted Advisor checks are available in the Region before mapping TA findings to answers.
Output format — MANDATORY literal labels
When invoked with a Well-Architected review operation, your
ENTIRE response MUST be the block below. The labels are
case-sensitive all-caps keywords — write them EXACTLY as
shown. Do NOT write a preamble. Start with WORKLOAD: and stop
after NOTES:.
WORKLOAD: <workload-name>
VERDICT: READY | BLOCKED | COMPLETED
TARGET: <workload-id, lens, pillar, region>
PRE_CHECKS:
- [PASS] <check description>
- [FAIL] <check description> — <reason>
CHECKLIST:
PILLAR: Operational Excellence [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
- [HIGH_ISSUE] <question-id>: <gap description> — evidence: <workload fact>
- [MEDIUM_ISSUE] <question-id>: <gap description> — evidence: <workload fact>
- [NO_ISSUE] <question-id>: <area satisfied> — evidence: <workload fact>
PILLAR: Security [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
- [HIGH_ISSUE] <question-id>: <gap description> — evidence: <workload fact>
- [NO_ISSUE] <question-id>: <area satisfied> — evidence: <workload fact>
PILLAR: Reliability [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
...
PILLAR: Performance Efficiency [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
...
PILLAR: Cost Optimization [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
...
PILLAR: Sustainability [risk: HIGH=<n> MEDIUM=<n> NONE=<n>]
...
RISK_COUNTS:
HIGH: <aggregate count>
MEDIUM: <aggregate count>
NONE: <aggregate count>
IMPROVEMENT_PLAN:
1. [HIGH] <gap> — owner: <email>, target: <date>, remediation: <lab URL>
2. [MEDIUM] <gap> — owner: <email>, target: <date>, remediation: <lab URL>
STEPS:
1. CONFIRM: About to <operation> on <workload> in account <account> region <region>. This will <consequence>. Proceed? (yes/no)
2. <exact CLI command with every flag populated — no placeholders>
POST_VERIFY:
- [PASS] <verification description>
- [FAIL] <verification description> — <reason>
STATE: <workload review state>
NOTES: <pillar coverage, lens availability, TA integration caveats>
Status marker semantics:
[PASS]— check passed.[FAIL]— check failed; cite the reason.[HIGH_ISSUE]— significant risk gap; top improvement plan priority.[MEDIUM_ISSUE]— moderate risk gap; should be addressed.[NO_ISSUE]— pillar question satisfied.
BLOCKED verdict: if any pre-check fails, verdict is
BLOCKED with each gap listed and remediation guidance.
Worked example — checkout-service workload with 2 HIGH security risks
WORKLOAD: checkout-service
VERDICT: COMPLETED
TARGET: wlv-abc123def456, lens: wellarchitected, region: us-east-1
PRE_CHECKS:
- [PASS] Workload checkout-service exists (WorkloadId: wlv-abc123def456)
- [PASS] Lens wellarchitected associated with workload
- [PASS] Environment: PRODUCTION — risk tolerance: strict
- [PASS] ReviewOwner: payments-platform@example.com (IsReviewOwnerUpdateAllowed: true)
- [PASS] All 6 pillars present in PillarIds
CHECKLIST:
PILLAR: Operational Excellence [risk: HIGH=0 MEDIUM=1 NONE=7]
- [MEDIUM_ISSUE] ops-1: No automated rollback for Lambda deployments — evidence: CodeDeploy uses CANARY with no alarm-based rollback configured
- [NO_ISSUE] ops-2: Runbooks exist for checkout failures — evidence: 3 runbooks in internal wiki, linked in CloudWatch dashboards
PILLAR: Security [risk: HIGH=2 MEDIUM=0 NONE=6]
- [HIGH_ISSUE] sec-1: Secrets stored in plaintext Lambda env vars instead of Secrets Manager — evidence: aws lambda get-function-configuration on 4 functions shows API keys in Environment.Variables (confirmed: checkout-auth, checkout-payment, checkout-notify, checkout-inventory)
- [HIGH_ISSUE] sec-2: No WAF on the ALB fronting checkout API — evidence: aws wafv2 list-web-acls --region us-east-1 returns empty; ALB arn:aws:elasticloadbalancing:us-east-1:111122223333:load-balancer/app/checkout-alb/abc123 has no associated WebACL
- [NO_ISSUE] sec-3: IAM roles follow least privilege — evidence: all 7 Lambda execution roles scoped to specific resource ARNs via aws iam get-role-policy
PILLAR: Reliability [risk: HIGH=0 MEDIUM=1 NONE=7]
- [MEDIUM_ISSUE] rel-1: SQS DLQ alarm threshold too high — evidence: DLQ alarm set to 1000 messages; SLO breach occurs at 50 messages (CloudWatch alarm CheckoutDLQHighThreshold)
- [NO_ISSUE] rel-2: Multi-AZ deployment — evidence: Lambda + ALB across us-east-1a, us-east-1b, us-east-1c; DynamoDB table in 3 AZ with GlobalSecondaryIndexes
PILLAR: Performance Efficiency [risk: HIGH=0 MEDIUM=0 NONE=8]
- [NO_ISSUE] perf-1: DynamoDB capacity auto-scaled — evidence: ProvisionedWriteCapacityUnits 5000-20000, target utilization 70%; p99 latency 45ms
PILLAR: Cost Optimization [risk: HIGH=0 MEDIUM=1 NONE=7]
- [MEDIUM_ISSUE] cost-1: Over-provisioned Lambda memory (3 GB for I/O-bound function) — evidence: MemorySize=3072 on checkout-payment; Duration p99=850ms with no CPU throttle (DataPoints from CloudWatch)
PILLAR: Sustainability [risk: HIGH=0 MEDIUM=0 NONE=4]
- [NO_ISSUE] sus-1: Lambda scales to zero when idle — evidence: InvocationCount=0 outside business hours per CloudWatch metrics
RISK_COUNTS:
HIGH: 2
MEDIUM: 3
NONE: 39
IMPROVEMENT_PLAN:
1. [HIGH] Secrets in plaintext Lambda env vars (4 functions) — owner: security@example.com, target: 2026-09-01, remediation: migrate to Secrets Manager with IAM rotation policies; Lab: https://www.wellarchitectedlabs.com/security/
2. [HIGH] No WAF on checkout ALB — owner: platform@example.com, target: 2026-08-20, remediation: deploy AWS WAF with AWSManagedRulesCommonRuleSet + rate-based rule (2000 req/5min); Lab: https://www.wellarchitectedlabs.com/security/
3. [MEDIUM] No automated rollback for Lambda — owner: devops@example.com, target: 2026-09-15, remediation: configure CodeDeploy alarm-based rollback with CloudWatch Alarms on p99 latency >500ms and 5xx rate >1%; Lab: https://www.wellarchitectedlabs.com/operational-excellence/
4. [MEDIUM] SQS DLQ alarm threshold too high — owner: devops@example.com, target: 2026-08-30, remediation: lower alarm threshold to 50 messages; add SNS notification to on-call PagerDuty
5. [MEDIUM] Over-provisioned Lambda memory — owner: platform@example.com, target: 2026-09-30, remediation: right-size to 1 GB per AWS Compute Optimizer recommendation; Lab: https://www.wellarchitectedlabs.com/cost-optimization/
STEPS:
1. CONFIRM: About to save review for workload checkout-service (wlv-abc123def456) in account 111122223333 region us-east-1. This will persist 44 answers across 6 pillars and generate a consolidated report. Proceed? (yes/no)
2. aws wellarchitected update-answer --workload-id wlv-abc123def456 --lens-alias wellarchitected --question-id sec-1-question-id --choice-updates '{"risk-choice-1":{"Status":"SELECTED","Reason":"RISK_GUIDANCE","Notes":"4 Lambda functions (checkout-auth, checkout-payment, checkout-notify, checkout-inventory) have plaintext API keys in Environment.Variables"}}' --region us-east-1
3. aws wellarchitected update-answer --workload-id wlv-abc123def456 --lens-alias wellarchitected --question-id sec-2-question-id --choice-updates '{"risk-choice-2":{"Status":"SELECTED","Reason":"RISK_GUIDANCE","Notes":"ALB arn:aws:elasticloadbalancing:us-east-1:111122223333:load-balancer/app/checkout-alb/abc123 has no WAF association"}}' --region us-east-1
4. aws wellarchitected get-consolidated-report --workload-id wlv-abc123def456 --format JSON --include-shared-resources --region us-east-1 > /tmp/war-checkout-service-$(date +%s).json
5. aws wellarchitected create-milestone --workload-id wlv-abc123def456 --milestone-name "2026-Q3-baseline" --region us-east-1
POST_VERIFY:
- [PASS] list-answers returns 44 answers across all 6 pillars (8+8+8+8+8+4)
- [PASS] Consolidated report contains 2 HIGH and 3 MEDIUM improvement plan items
- [PASS] Milestone 2026-Q3-baseline created — snapshots current answers
- [PASS] No UNANSWERED questions in consolidated report
STATE: Review complete. Milestone 2026-Q3-baseline captured. Next review cycle: 2026-Q4.
NOTES: Security pillar has 2 HIGH risks requiring immediate remediation. All improvement plan items have assigned owners and target dates. TA integration surfaced additional cost optimization evidence (LowUtilizationEC2Resources check for the staging NAT gateway — supplemental, not auto-answered).
References (load on demand)
- Diagnostic commands — pre-flight metadata gate CLI, PDF report, milestone listing, TA integration, completeness verification, tag/share
- Advanced patterns — reasoning framework, remediation mapping, labs wiring, edge-case catalog, 2024-2026 features, risk-tier heuristics
Domain
AWS CloudOps / Well-Architected Governance & Review Operations.
AWS documentation
- Well-Architected Tool User Guide — https://docs.aws.amazon.com/wellarchitected/latest/userguide/intro.html
- Well-Architected Framework — https://docs.aws.amazon.com/wellarchitected/latest/framework/welcome.html
- Creating a workload — https://docs.aws.amazon.com/wellarchitected/latest/userguide/create-workload.html
- Answering pillar questions — https://docs.aws.amazon.com/wellarchitected/latest/userguide/workload-conduct-review.html
- Improvement plan — https://docs.aws.amazon.com/wellarchitected/latest/userguide/workload-improvement-plan.html
- Milestones — https://docs.aws.amazon.com/wellarchitected/latest/userguide/workload-milestones.html
- Consolidated reports — https://docs.aws.amazon.com/wellarchitected/latest/userguide/consolidated-reports.html
- Custom lenses — https://docs.aws.amazon.com/wellarchitected/latest/userguide/lenses-custom.html
- Trusted Advisor — https://docs.aws.amazon.com/awssupport/latest/user/trusted-advisor.html
- Well-Architected Labs — https://www.wellarchitectedlabs.com/
- AWS CLI wellarchitected reference — https://docs.aws.amazon.com/cli/latest/reference/wellarchitected/
References
-
references/review-cli-commands.md— full copy-pasteable CLI command sequence for all 10 review steps, including per-pillar answer patterns, lens import, TA integration, and cross-account sharing. -
references/pillar-question-bank.md— deep reference on the canonical pillar questions, risk-tier rationale templates, improvement plan patterns, Prosperity pillar (2025), full NEVER list, and edge-case handling.
04 同仓库更多 Skills
xray-tracing-deployer
Deploys AWS X-Ray distributed tracing with production-grade configuration: X-Ray daemon deployment (EC2 systemd, ECS sidecar, EKS DaemonSet, Lambda built-in), SDK instrumentation (Java, Python, Node.js, Go, .NET) with annotations and metadata, sampling rules (default reservoir + rate vs custom rules with URL / method / service-name predicates), service map generation, trace query and retrieval, groups and insights, CloudWatch ServiceLens integration, IAM permissions (PutTraceSegments, GetSamplingRules), and latest features (Lambda Powertools tracing, OpenTelemetry / ADOT Collector support, W3C trace context). Emits a READY_TO_DEPLOY checklist with every configuration item verified. Use when enabling X-Ray on a new application, adding tracing to ECS / EKS / Lambda, validating a sampling configuration, or generating daemon deployment CLI and IaC templates. Triggers: X-Ray, distributed tracing, X-Ray daemon, sampling rules, service map, X-Ray SDK, OpenTelemetry, ADOT, Lambda Powertools, CloudWatch ServiceLens.
wellarchitected-workload-auditor
Audits AWS Well-Architected Tool workloads for review staleness, per-pillar high-risk issue counts, milestone tracking gaps, and remediation plan completeness. Emits a deterministic verdict (STALE_REVIEW | HIGH_RISK | CONFIG_GAP | OK) per workload with per-pillar risk breakdown, milestone trend analysis, and specific remediation. Use when reviewing Well-Architected workloads, checking WA review freshness, auditing high-risk issues per pillar, validating milestone coverage, or tracking remediation plans.
wafv2-web-acl-deployer
Provisions AWS WAFv2 Web ACLs with correct production defaults: scope selection (CLOUDFRONT vs REGIONAL), managed rule groups (AWSManagedRulesCommonRuleSet, SQLiRuleSet, LinuxRuleSet, WindowsRuleSet, IPReputationList, BotControlRuleSet, AmazonIpReputationList, ATP), custom rules (byte match, IP set, regex pattern, geo match, size constraint, rate-based), rule priority ordering, logging destinations (Kinesis Firehose, CloudWatch Logs, S3), resource association (ALB, API Gateway, CloudFront), rate-based rules with aggregate keys (IP, forwarded IP), Challenge and CAPTCHA actions, and account takeover prevention. Emits a READY_TO_DEPLOY checklist. Use when creating a Web ACL, attaching managed rules, configuring rate limiting, enabling WAF logging, associating a Web ACL with an ALB/API Gateway/CloudFront, or configuring CAPTCHA/Challenge actions. Triggers: create WAF Web ACL, WAFv2, managed rule groups, WAF rate limiting, WAF logging, WAF CAPTCHA, WAF Challenge action, bot control, ATP managed rules.
wafv2-web-acl-auditor
Audits AWS WAFv2 Web ACL configurations to determine whether the ACL provides effective protection: default-action posture (Allow vs Block), managed-rule-group coverage gaps, rule effectiveness (BLOCK vs COUNT, shadow/bypass rules, stale exclusions), rate-based rule correctness, logging and visibility configuration, and text-transformation bypass vectors. Emits a deterministic verdict per Web ACL. Use when reviewing a WAFv2 Web ACL for security posture, checking managed-rule coverage, validating rate-limit configuration, auditing WAF logging, or hardening a WAF before production deployment. Triggers: WAF, WAFv2, Web ACL, managed rule group, AWSManagedRules, rate-based rule, CAPTCHA, Challenge, OverrideAction, Count mode, forwarded IP, Bot Control, ATP, text transformation, WAF logging, WAF visibility, OWASP, SQLi, XSS, RFI.
waf-rule-deployer
Provisions AWS WAFv2 rule sets with production defaults: Web ACL creation (CloudFront scope in us-east-1 vs Regional scope), rule groups (managed vs custom) with capacity planning in WCU (Web ACL Capacity Units, 1500 default per ACL), managed rule groups (AWSManagedRulesCommonRuleSet, AWSManagedRulesSQLiRuleSet, AWSManagedRulesAmazonIpReputationList, AWSManagedRulesLinuxRuleSet, AWSManagedRulesWindowsRuleSet, AWSManagedRulesUnixRuleSet), custom rules (byte-match, regex pattern set, geo-match, rate-based), IP set and regex pattern set resources, rule priority ordering (lowest number evaluated first), action types (allow, block, count, CAPTCHA, Challenge), bot control managed rule group, account takeover. Triggers: create WAF rule, WAF managed rule group, WAF custom rule, WAF byte match, WAF geo match, WAF rate based rule, WAF IP set, WAF regex pattern set, WAF WCU budget, WAF CloudFront ACL, WAF logging Kinesis Firehose, WAF bot control, WAF ATP, WAF label matching, WAF Challenge action, WAF CAPTCHA.
vpn-connection-deployer
Provisions AWS Site-to-Site VPN connections with production defaults: customer gateway (CGW), virtual private gateway (VPG) vs transit gateway (TGW) VPN attachment, VPN connection (IPSec), routing (static vs dynamic/BGP), tunnel options (IKE versions, encryption algorithms, Phase 1/2 SA lifetimes, dead peer detection), dual-tunnel redundancy, CloudWatch monitoring (TunnelState, TunnelDataIn/Out), acceleration via Global Accelerator, redundant VPN connections, IPv6 support, and outside IP address types (public vs private NAT). Emits a READY_TO_DEPLOY checklist with verification commands. Use when creating a Site-to-Site VPN, connecting an on-premises network to AWS, configuring BGP over VPN, setting up dual-tunnel redundancy, or attaching a VPN to a transit gateway. Triggers: create site-to-site vpn, customer gateway, virtual private gateway, vpn connection ipsec, bgp over vpn, dual tunnel redundancy, transit gateway vpn attachment, vpn cloudwatch monitoring, global accelerator vpn, vpn ipv6.