aws-security
An agent skill by aws, from aws/agent-toolkit-for-aws. Tags: aws, compliance, security, threat-hunting, validation.
What it does
Covers AWS security services and workflows — Security Hub V2 (OCSF) findings, connectors, aggregators, automation rules, and security posture summaries; Security Hub CSPM (V1/ASFF) controls and compliance standards; GuardDuty threat findings; Inspector vulnerability findings; Macie sensitive data findings; Detective investigation; and Security Lake configuration and data aggregation. Applicable when questions involve security posture, Exposure findings, CSPM failed controls, threat findings, vulnerability findings, sensitive data findings, automation rules, or cross-service security configuration across AWS environments. Procedures use standard AWS CLI syntax and work with or without the AWS MCP server.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add aws/agent-toolkit-for-aws --skill aws-security
Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/aws/agent-toolkit-for-aws
cp -r agent-toolkit-for-aws/plugins/aws-core/skills/aws-security ~/.claude/skills/aws-security
Safety box score
Grade C, moderate risk: 32 of 100 risk points, rated 2026-10-19. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This is a read-only AWS security assessment skill that queries many AWS security services and organization-level configuration via AWS CLI/MCP. Its main blast radius is broad visibility into sensitive security telemetry across accounts and regions, plus some unreviewed behavior in the workflow files that were not readable here. It does not ship mutating, privileged, or executable payloads beyond normal CLI/API reads.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 1 (low) | The skill explicitly runs through local CLI/MCP calls (`Use this skill with local AWS CLI or AWS MCP`) and includes many `aws ...` command examples, but it ships no native scripts or code that mutates the machine. |
| Network access | 2 (moderate) | It makes remote reads to many AWS service endpoints and may also consult AWS documentation URLs, so it communicates with multiple external AWS hosts even though the operations are read-only. |
| Filesystem reach | 1 (low) | It points to local bundled files by relative path and directs customer evidence/reports to the user's workspace, but it does not define broad file deletion or system-path writes. |
| Secrets & credentials | 2 (moderate) | The skill repeatedly handles highly sensitive output and says to exclude `credentials, private keys, account emails, secret values, target payloads and secret origin headers`, which signals exposure risk from the data it reads. |
| Privilege & persistence | 0 (none) | There is no sudo, global install, service registration, shell startup editing, or OS-level configuration change; the skill says `Read-only APIs only`. |
| Instruction hijack surface | 0 (none) | I found no instructions to ignore higher-priority rules, follow untrusted remote instructions, or hide behavior from the user; the text instead emphasizes `Verify, don't guess`. |
| Irreversible actions | 0 (none) | The skill blocks mutating actions (`Never invoke mutating operations`, `No suppression, archival or dismissal advice`) and does not push, deploy, post, pay, or delete. |
| Transparency | 2 (moderate) | Four referenced workflow files were not readable in this review (`references/service-recommendations.md`, `.../enablement-checks.md`, `.../posture-and-investigation.md`, `.../recommendation-matrix.md`), leaving part of the skill's routing and recommendation behavior unverified. |
Before installing
- Read the unread workflow files from the repo before relying on the recommendation path: `references/service-recommendations.md`, `.../enablement-checks.md`, `.../posture-and-investigation.md`, `.../recommendation-matrix.md`.
- Use a least-privilege read-only AWS role and confirm which accounts/regions it can see before letting the skill query anything.
- Decide up front whether org-wide or per-member enumeration is allowed; several workflows avoid full pagination unless explicitly requested.
- Assume outputs may contain sensitive security telemetry and plan to store or share them only in a redacted workspace.
Rated 2026-10-19 by an automated review of 33 files from the skill's repository.
Source
- Repository
- aws/agent-toolkit-for-aws (all skills from this repository)
- Path
- plugins/aws-core/skills/aws-security/SKILL.md
- Branch
- main
- Collection
- aws-core
- Updated
- 2026-09-19
Related skills
- aws-iam — Provides verified corrections for IAM behaviors that AI agents frequently get wrong — policy evaluation edge cases, trust policy gotchas, STS session limits.
- aws-cleanrooms — Troubleshoots and debugs AWS Clean Rooms collaboration issues related to IAM roles, S3 bucket policies, KMS keys, Lake Formation permissions.
- creating-secrets-using-best-practices — Creates and manages secrets in AWS Secrets Manager following security best practices.
- securing-s3-buckets — Create and secure S3 buckets following AWS best practices for access control, encryption, monitoring, and remediation of misconfigurations.
- setting-up-cloudtrail-multi-region — Enables a multi-region AWS CloudTrail trail with S3 log storage, CloudWatch Logs integration.
- aws-cloudformation — Authors, validates, and troubleshoots AWS CloudFormation templates.