agentleFS
Sign inStart free
kaandgnshkaandgnsh/Claude-Blue/skills/aws-security/aws-iam-review/SKILL.md

aws-iam-review

Skill0 starsChanged 14 days agoMITView on GitHub

Review AWS IAM for over-privilege and escalation paths — admin/wildcard policies, PassRole and other escalation actions, risky trust policies (external accounts, wildcard principals, weak OIDC conditions), stale users and keys, root usage, unused access — with IAM Access Analyzer, credential reports and a bundled tested policy linter. Use for AWS account security reviews, least-privilege projects, or after credential compromise.

The file

Create a free account to mount
Files
SKILL.md

AWS IAM Review

Collect

aws iam get-account-authorization-details > auth.json          # all users, roles, groups, policies (incl. trust)
python3 scripts/iam_policy_lint.py auth.json --min-severity medium   # bundled linter (exit 1 on critical)
aws iam generate-credential-report && aws iam get-credential-report --query Content --output text | base64 -d > cred.csv
aws accessanalyzer list-findings --analyzer-arn <external-access-analyzer-arn> --filter '{"status":{"eq":["ACTIVE"]}}'
aws accessanalyzer validate-policy --policy-type IDENTITY_POLICY --policy-document file://policy.json

IAM Access Analyzer: external access analyzer (resources shared outside the zone of trust), unused access analyzer (unused roles, keys, permissions), and policy generation from CloudTrail activity for least privilege.

Credential report checks (cred.csv)

  • root: access_key_1_active must be false; MFA on; password_last_used recent = investigate.
  • Users with password_enabled and no mfa_active.
  • Access keys older than 90 days (access_key_1_last_rotated) or unused (access_key_1_last_used_date N/A).
  • Users with console + keys — prefer SSO (IAM Identity Center) for humans; no long-term keys for humans.

Trust policy red flags

  • "Principal": {"AWS": "*"} or "*" without strict conditions.
  • Cross-account trust to unknown accounts; third-party roles without sts:ExternalId condition (confused deputy).
  • OIDC federation (GitHub Actions, EKS IRSA) without sub restriction — e.g. token.actions.githubusercontent.com:sub must name repo and branch/environment.
  • SAML providers you don't recognize.

Escalation actions (if granted broadly)

iam:PassRole with compute creation (ec2:RunInstances, lambda:CreateFunction, glue:CreateDevEndpoint, cloudformation:CreateStack), iam:CreatePolicyVersion, iam:Attach*/Put*Policy, iam:CreateAccessKey on other users, iam:UpdateAssumeRolePolicy, ssm:SendCommand/StartSession on privileged instances, lambda:UpdateFunctionCode on privileged functions.

Output

Principal | Finding | Evidence (policy/statement) | Severity | Fix | Owner plus metrics: users without MFA, keys > 90 days, admin principals count, external trusts.

Related skills

aws-org-guardrails, aws-credential-compromise, cloudtrail-investigation, privileged-access-review

References

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free