Skip to content
SkillModel-led3,238 tokens in use

hunt-cloud-misconfig

A hunting guide for cloud storage and compute misconfigurations across AWS, GCP and Azure, with a dedicated section on AWS CloudWatch RUM abuse.

14,533 stars on the repository MIT Licenseelementalsouls/Claude-BugHunter · skills/hunt-cloud-misconfig
/plugin marketplace add elementalsouls/Claude-BugHunter /plugin install claude-bughunter@elementalsouls
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing hunt-cloud-misconfig costs

199tokens idle
3,238tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

shelldetectednetworkdetectedcredentialsdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter12/15
  • Structure17/25
  • Examples12/15
  • Output15/15
  • Size11/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Good
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 13
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 13
  • Reads credentials−15

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 35

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity100
  • Maintenance84
  • Community68
  • Maturity28
  • Momentum40

hunt-cloud-misconfig is a skill for Claude covering cloud-native storage and compute misconfiguration: public S3, GCS and Azure Blob buckets, open Firebase rules, EC2 metadata reachable through SSRF, exposed admin panels and message brokers, and leaked IAM credentials in JavaScript bundles. Kubernetes and Docker exposure, CI/CD pipeline attacks, and post-credential IAM escalation are each owned by a separate skill in the same bundle.

The skill gives detection commands for bucket listing and common-name guessing across AWS, GCP and Azure, a local-verification workflow using LocalStack, and an extended section on AWS CloudWatch RUM: how to fingerprint the identity-pool and guest-role configuration a RUM snippet exposes by design, how to extract AWS credentials from an over-permissioned guest role, and a severity rubric tied to what that role grants. It is for assets covered by a program or engagement you are authorized to test.

What You Get

  • Detection commands for public S3, GCS and Azure Blob buckets, common-bucket-name guessing, and open Firebase read/write rules
  • A method for reaching EC2 instance metadata credentials through SSRF, and a list of exposed admin panel and message-broker paths to probe
  • A LocalStack-based local-verification workflow with the environment variables needed to work around a known awscli compatibility issue
  • A full AWS CloudWatch RUM section: JS bundle fingerprints, a regex set for extracting the identity pool ID and guest role ARN, and four attack chains from credential extraction to dashboard poisoning
  • A validation checklist and severity rubric for the RUM finding tied to exactly which AWS actions the exposed guest role grants beyond intended telemetry use

Common Use Cases

  • Scanning a target’s JavaScript bundles and common bucket-name patterns for a publicly readable or writable S3, GCS or Azure Blob bucket
  • Testing whether an SSRF-reachable endpoint exposes EC2 instance metadata credentials
  • Extracting and validating AWS credentials from an over-permissioned Cognito guest role exposed by an embedded AWS CloudWatch RUM snippet

What Makes It Unique

The skill goes well past the standard bucket-listing checklist with a full treatment of AWS CloudWatch RUM, a newer and less-documented attack surface where the identity-pool ID and guest-role ARN are public by design and the vulnerability is scoped entirely to how broad the guest role’s IAM policy is. It backs this section with named research from Rhino Security Labs, Datadog Security Labs and HackTricks Cloud, and a validation checklist that requires at least one allowed action beyond the documented telemetry permission before the finding counts as more than informational.

Under The Hood

The skill is a single SKILL.md of guidance with no bundled scripts. Following it, the agent sends HTTP requests to cloud storage endpoints, Firebase realtime-database URLs, and common admin-panel paths, greps JavaScript bundles for RUM configuration fields and AWS credential patterns, and runs AWS CLI commands such as cognito-identity get-credentials-for-identity and sts get-caller-identity to validate extracted credentials. It also documents a Docker-based LocalStack setup for local verification before testing against real cloud infrastructure. Use it only against assets you are authorized to test.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers