AGENT SKILL

service-quotas-usage-auditor

Audits AWS Service Quotas — quota utilization per service, approaching limits (>=80%), CloudWatch alarm coverage on AWS/Usage metrics, applied vs default quota drift, adjustable quotas stuck at default with rising usage, denied or stale quota increase requests, and non-trackable quotas lacking a UsageMetric. Emits a deterministic verdict (APPROACHING_LIMIT | NO_ALARM | CONFIG_GAP | OK) per quota with enumerated findings and CLI remediation. Use when reviewing quota utilization, checking for approaching service limits, auditing quota increase request history, or validating CloudWatch alarm coverage.

License 许可
Apache-2.0
最近更新
2026-09-24
Stars
★ 4
Forks
0

01 前言 Front Matter

name
service-quotas-usage-auditor
license
Apache-2.0
metadata
{ "tags": "service-quotas, management, quota, limits, cloudwatch, monitoring, audit", "phase": "2", "author": "Jacky Chan — AWS Community Builder", "domain": "aws-cloudops", "family": "Management", "version": "0.1.0", "keywords": "Service Quotas, quota utilization, approaching limit, AWS/Usage, CloudWatch alarm, quota increase, applied quota, default quota, Adjustable, UsageMetric, quota code, service limits, Trusted Advisor, quota monitoring, quota remediation", "complexity": "medium", "entry_point": "false", "when_to_use": "Auditing AWS Service Quotas utilization, CloudWatch alarm coverage on AWS/Usage metrics, and quota increase request history.", "requires_llm": "true", "verdict_shape": "APPROACHING_LIMIT | NO_ALARM | CONFIG_GAP | OK", "invocation_schema": "Input: either (a) a Service Quotas snapshot (quota code, service code, applied value, default value, UsageMetric, utilization, increase request history, CloudWatch alarm state), OR (b) a service-code for live-account audit via aws service-quotas list-service-quotas. Output: deterministic QUOTA/VERDICT/REASON/UTILIZATION/FINDINGS/ REMEDIATION block per quota, where VERDICT is one of APPROACHING_LIMIT, NO_ALARM, CONFIG_GAP, OK (or ERROR for malformed input).", "supports_pipeline": "true", "activation_triggers": "audit service quotas, check quota utilization, approaching service limit, quota increase request, cloudwatch alarm on quota, service limits audit" }
description
Audits AWS Service Quotas — quota utilization per service, approaching limits (>=80%), CloudWatch alarm coverage on AWS/Usage metrics, applied vs default quota drift, adjustable quotas stuck at default with rising usage, denied or stale quota increase requests, and non-trackable quotas lacking a UsageMetric. Emits a deterministic verdict (APPROACHING_LIMIT | NO_ALARM | CONFIG_GAP | OK) per quota with enumerated findings and CLI remediation. Use when reviewing quota utilization, checking for approaching service limits, auditing quota increase request history, or validating CloudWatch alarm coverage.
compatibility
Agent runtime that reads SKILL.md (Claude Code, Cursor, Windsurf, Codex, Gemini). No AWS CLI required for offline quota-snapshot classification. Live-account audits use aws service-quotas list-service-quotas, get-service-quota, get-aws-default-service-quota, list-requested-service-quota-change-history, and aws cloudwatch describe-alarbs-for-metric (AWS CLI v2, SSO or key-based credentials).

02 资源文件 18

03 SKILL.md 正文

Service Quotas Usage Auditor

Mindset

One-line takeaway: the verdict is the worst finding across all dimensions, and three Service Quotas behaviors are easy to misjudge — applied vs default (list-service-quotas returns the APPLIED value, not the AWS default), UsageMetric presence (quotas without a metric block CANNOT be auto-monitored via CloudWatch), and request status lifecycle (a PENDING increase has not taken effect — the applied quota is still the old value until APPROVED).

A quota is the hard ceiling on a specific AWS resource or API rate. Exhausting one is a silent outage: API calls return ThrottlingException or LimitExceededException with no remediation window. Unlike a misconfigured policy (fixable by editing text), a quota that blocks operations can stall an entire deployment pipeline while a support case grinds through the approval process.

Quick reference — verdict thresholds

Condition Verdict Step
Utilization >= 80% of applied quota (inclusive: exactly 80% counts) APPROACHING_LIMIT 1
Adjustable: false AND utilization >= 50% CONFIG_GAP 2b
Increase request DENIED/NOT_APPROVED AND utilization >= 50% CONFIG_GAP 2c
At AWS default, Adjustable: true, utilization >= 50%, request history completely empty CONFIG_GAP 2d
No UsageMetric block defined CONFIG_GAP 2a
Has UsageMetric, utilization < 80%, no CloudWatch alarm NO_ALARM 3
Utilization < 50%, alarm in place or N/A OK 4

See the ordered steps below for edge cases. Deep Service Quotas + CloudWatch integration details are in the Deep reference section at the end.

Pre-flight: quota metadata gate (run before classification)

Before evaluating utilization, classify the quota metadata itself. Several attributes short-circuit the audit — misclassifying them produces false positives.

Attribute Value Effect on audit
UsageMetric absent No CloudWatch metric exists. Cannot auto-monitor utilization. This is a CONFIG_GAP (Step 2a) regardless of current utilization — you are flying blind.
UsageMetric present Proceed — utilization is auto-trackable via AWS/Usage.
Adjustable false Hard limit. Cannot request an increase via request-service-quota-increase. If utilization >= 50%, the only remediation is architectural (distribute load, use a different service, reduce usage). Flag as CONFIG_GAP (Step 2b).
Adjustable true Quota is increasable. Proceed to check increase request history.
GlobalQuota true Account-wide quota (e.g., IAM users, S3 buckets). Applies across all regions. An increase applies globally.
GlobalQuota false Regional quota (e.g., VPCs per region, EC2 vCPUs per region). An increase in us-east-1 does NOT apply to eu-west-1 — request per-region.
Unit Gigabytes, Mbps, etc. Unit mismatch risk. Never compare a utilization value in one unit against a quota in another. A quota of "100 Gigabytes" at "50 Count" usage is a data error.
Applied == Default — Quota has never been increased. Not inherently bad, but combined with utilization >= 50% and Adjustable: true, it signals that no proactive capacity management has occurred (Step 2d).
Applied > Default — An increase was APPROVED at some point — active capacity management. This is a positive signal.

Multi-service / account-wide sweep note (pagination): when auditing every quota for a service, aws service-quotas list-service-quotas --service-code <code> returns at most 100 per page. Use --starting-token from the prior NextToken to page through all quotas. The same applies to list-requested-service-quota-change-history. Always drain the paginator to completion — the long tail often contains the quota that matters.

Malformed input handling. If the quota snapshot is missing the applied value, missing quota code, has a utilization value not parseable as a number, or has a unit mismatch between utilization and quota, output an ERROR block and refuse classification. Do NOT guess the utilization percentage — a wrong percentage produces a wrong verdict. Output:

QUOTA: <quota-code or name>
VERDICT: ERROR
REASON: Quota snapshot is incomplete or malformed — cannot classify.
REMEDIATION: Retrieve the canonical data with `aws service-quotas
get-service-quota --service-code <code> --quota-code <code> --output json` and re-audit.

Process — Classification logic (apply in order, aggregate worst)

Step 0: Expert knowledge — non-obvious Service Quotas behaviors

These behaviors are easy to misjudge without operational Service Quotas experience. Each changes a verdict if ignored:

Moved to references/advanced-patterns.md. Fifteen diagnosis-changing behaviors: applied-vs-default, UsageMetric presence, exact CW dimensions, GlobalQuota scope, pending-vs-applied, request lifecycle, Adjustable:false, ingestion delay, no utilization field, API throttling, auto-case routing, pending-request quota, QuotaCode stability, Trusted Advisor deprecation, instant spikes.

Step 1: Utilization threshold — APPROACHING_LIMIT

Compute utilization: utilization_pct = (currentUsage / appliedValue) * 100.

If utilization_pct >= 80, the verdict is APPROACHING_LIMIT. This is the highest-priority finding regardless of alarm state, UsageMetric presence, or increase request status. The operational risk of imminent exhaustion dominates all monitoring and structural concerns.

Boundary rule (critical): the threshold is inclusive — >= 80% means a quota at exactly 80.0% utilization IS APPROACHING_LIMIT, not OK. For example, 40/50 VPCs (exactly 80%) is APPROACHING_LIMIT. Models and operators frequently misread >= as > (strictly greater than); this is the single most common classification error. When in doubt, a quota at the exact boundary has no headroom for even one additional resource — treat it as approaching.

  • At exactly 80%: no headroom remains for growth. APPROACHING_LIMIT.
  • At 80-89%: immediate action required. Request an increase if adjustable, or reduce usage.
  • At 90-100%: critical — any new resource creation or API burst may fail with LimitExceededException.
  • At > 100%: the quota has been exceeded. AWS may allow brief overages for some quotas but will block new operations imminently.

If a PENDING increase request exists, note it in FINDINGS but do NOT downgrade the verdict — the applied quota has not changed yet.

If an APPROVED increase exists that has not yet propagated (applied value still shows old number), note the expected new value in FINDINGS.

Step 2: Structural config gap — CONFIG_GAP

If Step 1 did not match (utilization < 80%), check for structural issues in this order:

(a) No UsageMetric defined. The quota cannot be auto-monitored via CloudWatch. This is a CONFIG_GAP regardless of utilization — you have no programmatic signal to detect exhaustion. The remediation is to track manually (enumerate live resources and compare against the applied value on a schedule) or use AWS Config rules.

(b) Adjustable: false AND utilization >= 50%. The quota is a hard limit that cannot be increased, and utilization is high enough to be a future risk. This is a CONFIG_GAP because there is no increase path — the remediation is architectural.

(c) Increase request DENIED or NOT_APPROVED AND utilization >= 50%. A previous increase attempt was rejected and utilization is moderate-to-high. The team has not addressed the underlying capacity issue. This is a CONFIG_GAP — either restructure usage or open a support case with justification.

(d) At AWS default, Adjustable: true, utilization >= 50%, AND the increase request history is completely empty (zero entries of ANY status — no PENDING, no APPROVED, no DENIED, no NOT_APPROVED). The quota has never been increased and no request has ever been submitted, but utilization is high enough to warrant proactive action.

Disambiguation (critical for D8): Step 2d matches ONLY when the request history list is truly empty. If there is ANY entry in the history — even a PENDING request or a DENIED request — Step 2d does NOT apply. A PENDING request means the team is actively seeking an increase (skip to Step 3). A DENIED request is handled by Step 2c. Only a completely empty history combined with utilization >= 50% at the AWS default value triggers this CONFIG_GAP.

If none of (a)-(d) match, proceed to Step 3.

Step 3: Alarm coverage — NO_ALARM

If the quota has a UsageMetric (auto-trackable), utilization is < 80%, and Step 2 found no structural issues, check for a CloudWatch alarm.

If NO CloudWatch alarm is configured on the AWS/Usage metric for this quota, the verdict is NO_ALARM. The quota is trackable but no alerting exists — a sudden utilization spike would go undetected until operations fail.

Verification: aws cloudwatch describe-alarms --metric-name <MetricName> --namespace AWS/Usage --dimensions <dims from UsageMetric>. An alarm is considered to exist ONLY when ALL of these conditions are met: (1) MetricName matches the UsageMetric.MetricName value, (2) every dimension from the UsageMetric.Dimensions list appears in the alarm with the same name AND value, and (3) the alarm Threshold is at or below the applied quota value — an alarm threshold of 999999 for a quota of 50 is effectively disabled and should be treated as NO_ALARM. An empty describe-alarms response or any dimension mismatch means NO_ALARM.

Note: this step is only reached if Step 2 did not match. A quota at default with 60% utilization is CONFIG_GAP (Step 2d), not NO_ALARM — the structural issue (should have requested an increase) takes precedence over the monitoring gap (no alarm).

Step 4: OK

If none of Steps 1-3 matched, the verdict is OK:

  • Utilization is below 80% (no approaching-limit risk).
  • No structural issues (Step 2 clean).
  • Either a CloudWatch alarm is configured, or the quota is low-risk (low utilization, not adjustable, no metric).

No remediation required. Recommend periodic re-audit as usage grows.

Step 5: Aggregation — worst finding wins

The final verdict is the maximum severity across all findings, where APPROACHING_LIMIT > CONFIG_GAP > NO_ALARM > OK:

verdict = max(utilization_severity, config_gap_severity, alarm_severity)

If no findings (all dimensions clean), the verdict is OK.

Creative audit extensions: The four verdicts cover the standard audit dimensions. You MAY add supplementary FINDINGS lines for quota-specific risks the standard steps do not capture (e.g., a quota shared across instance types where one type is near exhaustion, or a quota whose growth rate has accelerated in the past 30 days). Keep the VERDICT to the four standard values — additional context goes in FINDINGS, not in new verdict labels.

Output format (per quota)

QUOTA: <quota-code> (<quota-name>)
SERVICE: <service-code>
VERDICT: APPROACHING_LIMIT | NO_ALARM | CONFIG_GAP | OK
REASON: <1-2 sentences citing the worst finding and its step number>
UTILIZATION: <current>/<applied> (<pct>%) [default: <default>] [unit: <unit>]
FINDINGS:
  - [APPROACHING_LIMIT] <finding description (Step N)>
  - [CONFIG_GAP] <finding description (Step N)>
REMEDIATION: <specific CLI per finding, or "None required" if OK>

Worked example — approaching limit with denied increase

QUOTA: L-1212C26A (Running On-Demand Standard instances)
SERVICE: ec2
VERDICT: APPROACHING_LIMIT
REASON: Utilization is at 85% of the applied quota (850/1000 vCPUs).
A prior increase request to 2000 vCPUs was DENIED — the applied value
remains 1000 (Step 1).
UTILIZATION: 850/1000 (85%) [default: 1000] [unit: vCPUs]
FINDINGS:
  - [APPROACHING_LIMIT] 850/1000 vCPUs (85%) — any new instance launch may fail (Step 1)
  - [CONFIG_GAP] Increase request to 2000 DENIED on 2025-06-15 — no alternative capacity plan in place (Step 2c)
REMEDIATION:
  1. Open an AWS support case with workload justification for a vCPU increase.
  2. Switch non-critical workloads to Spot Instances (separate quota).
  3. Distribute across regions — the us-east-1 quota does not apply to us-west-2.

Anti-Patterns — NEVER

  • NEVER compute utilization against the AWS default value instead of the applied value. list-service-quotas returns the APPLIED value (which may include approved increases). Using the default inflates utilization and produces false-positive APPROACHING_LIMIT verdicts.

  • NEVER assume all quotas have a CloudWatch metric. Quotas without a UsageMetric block CANNOT be auto-monitored. There is no AWS/Usage metric to alarm on — the only option is manual resource enumeration. Treating "no alarm" as NO_ALARM for a non-trackable quota is a category error; it is CONFIG_GAP (Step 2a).

  • NEVER treat a PENDING increase request as if it has taken effect. The applied quota remains the old value until the request transitions to APPROVED. Computing utilization against the requested future value masks the current risk. Always use get-service-quota for the authoritative applied value.

  • NEVER suggest request-service-quota-increase on an Adjustable: false quota. It returns ValidationException. These are hard limits — the only remediation is architectural (redistribute, reduce, or switch services).

  • NEVER create a CloudWatch alarm on AWS/Usage without copying the exact dimensions from the UsageMetric block. Wrong dimension values produce an alarm that monitors nothing — DescribeAlarms shows OK while the actual metric has no data. Copy Service, Resource, Type, and any Class dimension verbatim.

  • NEVER assume a regional quota increase applies across regions. A VPC quota increase in us-east-1 does NOT apply to eu-west-1. Each region has independent quotas for non-global quotas. Request increases per-region.

  • NEVER use QuotaName as a stable identifier in automation. QuotaName can be renamed by the service team; QuotaCode (e.g., L-1212C26A) is the stable identifier. Scripts that key on QuotaName break silently when AWS renames a quota.

  • NEVER compare utilization across different Unit values. A quota of "100 Gigabytes" and usage of "50 Count" is a data error, not 50% utilization. Verify the unit before computing the ratio.

  • NEVER classify a quota at exactly 80.0% utilization as OK, NO_ALARM, or CONFIG_GAP. The >= 80% threshold is inclusive — 40/50 (exactly 80%) IS APPROACHING_LIMIT. The most common classification error is treating the boundary as exclusive (> 80%). If utilization computes to exactly 80.0%, the verdict MUST be APPROACHING_LIMIT, not OK.

  • NEVER ignore the 50% threshold for CONFIG_GAP. A quota at 55% utilization that is adjustable, at default, and has no increase request is a CONFIG_GAP — the team should have acted proactively. Waiting until 80% to flag it (where it becomes APPROACHING_LIMIT) leaves no remediation window.

  • NEVER recommend deleting resources as the only remediation for an APPROACHING_LIMIT quota. First check if the quota is adjustable and submit an increase request. Resource cleanup is a complement, not a substitute — the underlying capacity ceiling remains.

  • NEVER rely on Trusted Advisor service-limit checks as the sole utilization signal. Trusted Advisor covers only a subset of quotas and is deprecated for Business/Enterprise support tiers. The modern approach is Service Quotas + CloudWatch alarms on AWS/Usage.

  • NEVER drain only the first page of list-service-quotas (100 per page). The quota you need may be on page 2. Always iterate --starting-token to completion.

  • NEVER exceed the concurrent pending-request quota. Each account has a fixed limit on simultaneous PENDING increase requests per region (typically 4 per region per service). When this limit is full, request-service-quota-increase fails with QuotaExceededException even though the target quota has ample headroom. Before submitting a new request, check for stale PENDING entries via list-requested-service-quota-change-history --status PENDING and contact AWS support to withdraw expired ones — there is no public API to programmatically close a pending request.

  • NEVER classify a quota with UsageMetric present, an alarm configured, low utilization, and no structural issues as anything other than OK. Over-classifying clean quotas erodes trust in the audit.

Pre-flight safety checks (run before any remediation CLI)

  • MANDATORY CONFIRMATION GATE. Before any state-changing operation (request-service-quota-increase, closing a request), the auditor MUST emit: CONFIRM: About to request a quota increase for <quota-code> on <service-code> from <current> to <desired>. This opens a request that may require AWS support review. Proceed? (yes/no) Do NOT execute the CLI command until the operator confirms.
  • Before requesting an increase, verify the quota is Adjustable: true. A request on a non-adjustable quota returns ValidationException.
  • Before requesting an increase, verify --desired-value is strictly greater than the current applied value (get-service-quota). Equal or lower values are rejected.
  • Check for existing PENDING requests on the same quota to avoid duplicates: aws service-quotas list-requested-service-quota-change-history-by-quota --service-code <code> --quota-code <code --status PENDING.
  • Before creating a CloudWatch alarm, verify the UsageMetric dimensions by querying the metric first: aws cloudwatch get-metric-statistics --namespace AWS/Usage --metric-name <MetricName> --dimensions <dims> --start-time <iso> --end-time <iso> --period 300 --statistics <StatisticType>. No data points = wrong dimensions.
  • For regional quotas, confirm the correct --region is set on every CLI call. A quota increase in the wrong region is a waste of a request slot and cannot be transferred.

Remediation guidance

For APPROACHING_LIMIT

  1. If Adjustable: true, request an increase immediately: aws service-quotas request-service-quota-increase --service-code <code> --quota-code <code> --desired-value <value> Use a desired value with at least 50% headroom above current usage (e.g., if at 850/1000, request 1500-2000).
  2. If a PENDING request already exists, note the expected resolution time. Do NOT submit a duplicate — it wastes a request slot.
  3. If Adjustable: false, reduce usage immediately: terminate non-critical resources, switch to Spot/alternative instance types, redistribute across accounts or regions.
  4. Create or verify a CloudWatch alarm at 80% and 90% thresholds.

For CONFIG_GAP

  • No UsageMetric (Step 2a): set up a scheduled audit (EventBridge + Lambda) that enumerates live resources and compares against the applied quota. There is no CloudWatch metric to alarm on.
  • Adjustable: false (Step 2b): document the architectural constraint and plan capacity around it. Consider multi-account or multi-region distribution.
  • Denied increase (Step 2c): open an AWS support case with workload justification. Reduce usage in the interim.
  • At default, no request (Step 2d): submit a proactive increase request. Even if current usage is moderate, growth will eventually exhaust the default.

For NO_ALARM

Create a CloudWatch alarm on the AWS/Usage metric:

aws cloudwatch put-metric-alarm \
  --alarm-name "quota-<quota-code>-80pct" \
  --namespace AWS/Usage \
  --metric-name <MetricName from UsageMetric> \
  --dimensions <dims from UsageMetric> \
  --statistic <StatisticType from UsageMetric> \
  --period 300 \
  --threshold <applied * 0.8> \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 1 \
  --alarm-actions <sns-topic-arn>

For OK

No remediation required. Recommend:

  • Periodic re-audit as usage grows (monthly or per-deploy).
  • Verify the CloudWatch alarm threshold is still appropriate if the applied quota changes.
  • For regional quotas, audit each region where the workload operates.

Reference: AWS/Usage metric structure

The UsageMetric block maps to a CloudWatch metric in AWS/Usage:

Field Typical value Notes
MetricNamespace AWS/Usage Always this value.
MetricName ResourceCount Sometimes ResourceLimit or service-specific.
Dimensions Service, Resource, Type, optionally Class Copy verbatim — wrong values produce a silent alarm.
StatisticType Sum or Maximum Sum for cumulative resources; Maximum for peak-based quotas.

Organizations note: the management account can view member-account quotas via list-service-quotas --account-id <id> but CANNOT submit increase requests on their behalf — each member account must submit its own request.

Recent AWS features (2024-2026)

Moved to references/advanced-patterns.md. Organization-level quota management, more trackable quotas, increase-request automation.

References (load on demand)

  • advanced-patterns — Step 0 expert knowledge (15 non-obvious Service Quotas behaviors) and recent AWS features

Domain

AWS CloudOps / Service Quotas Capacity Management & Monitoring.

AWS documentation

04 同仓库更多 Skills

jhkchan
jhkchan/aws-agent-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.

★ 4
jhkchan
jhkchan/aws-agent-skills

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.

★ 4
jhkchan
jhkchan/aws-agent-skills

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.

★ 4
jhkchan
jhkchan/aws-agent-skills

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.

★ 4
jhkchan
jhkchan/aws-agent-skills

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.

★ 4
jhkchan
jhkchan/aws-agent-skills

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.

★ 4