Skip to main content

Local-first, read-only AWS audit CLI. Generate a VP/CFO-ready AWS audit report in minutes.

Project description

Kulshan

The blood test for your AWS bill.

Generate a local AWS audit report in minutes.

pip install kulshan
kulshan report

No setup. No data uploads. No infrastructure changes.

Just your AWS account and your laptop.

PyPI License Python

Documentation | IAM Policy | GitHub


What is Kulshan?

Kulshan reads your AWS account (Cost Explorer and your own CUR/Data Export files) and generates a business-ready report covering:

  • Cost anomalies and trends
  • Waste and orphaned resources
  • Tag compliance and cost attribution
  • Commitment health (RI/SP coverage)
  • Spend forecasting and acceleration
  • Security posture
  • DR readiness

Reads Cost Explorer and your own CUR data in place. No data leaves your environment.

Think of it as a baseline before deeper FinOps work, platform evaluations, or leadership reviews.


What You Get

An HTML report you can open in a browser and hand to your VP, CFO, or platform team. Also available as JSON, SARIF, and CSV.

The report scores your account 0-100 across each dimension, highlights the top findings by dollar impact, and provides an executive summary paragraph.


Install

pip install kulshan

Requires Python 3.9+. macOS, Linux, Windows.


AWS Credentials

Kulshan uses your existing AWS CLI credentials. If aws sts get-caller-identity works, Kulshan works.

The recommended flow is:

aws login
kulshan report

Kulshan identifies your active AWS identity automatically and routes to the correct local database.

Named profiles and explicit credentials also work:

# Named profile
kulshan --profile your-profile report

# Environment variable
AWS_PROFILE=customer-a kulshan report

# Assume role
kulshan --role-arn arn:aws:iam::123456789012:role/KulshanAudit report

Multi-Environment Isolation

Kulshan automatically separates data for different AWS identities.

On first use, Kulshan creates a local environment named after your AWS role or user:

✓ Created environment readonlyrole-cedar
  Using readonlyrole-cedar · account 1234…5678

Future runs with the same identity reuse the same environment and database. Different identities get separate environments automatically.

No --workspace flags or manual setup required.

Payer Binding

When Kulshan reads CUR data containing a bill_payer_account_id, it binds the environment to that payer:

✓ Bound this environment to payer XXXX-XXXX-9999 using CUR evidence.

Reconciliation

If multiple AWS identities access the same payer account, Kulshan offers to link them:

kulshan workspace reconcile

After linking, all identities route to one payer environment, and kulshan history shows a unified timeline.

Workspace Commands

kulshan workspace list                    # Show all environments
kulshan workspace show                    # Show current environment details
kulshan workspace rename ws_abc "Acme"    # Change display name
kulshan workspace reconcile               # Link shared-payer environments
kulshan workspace use ws_abc              # Set active workspace

History

kulshan history                    # Show scans (including linked environments)
kulshan history --direct-only      # Only scans from this workspace
kulshan history --account 123456789012   # Filter by credential account

Consolidated Reports

When a payer environment has multiple approved connections:

kulshan report     # Runs all approved connections, produces one report

The report shows coverage and source attribution:

Report status: Complete
Payer cost coverage: Verified payer-wide
Connections: billing, audit
Accounts verified: XXXX-XXXX-1111, XXXX-XXXX-2222

To run a single connection:

kulshan --connection audit report

Quick Commands

kulshan --version                       # Show version
kulshan doctor                          # Verify credentials and permissions
kulshan report                          # Default FinOps baseline (cost pack)
kulshan report --quick                  # Fast scan (skips confirmation)
kulshan report -o report.html           # Save as HTML
kulshan report --packs security,sweep   # Run specific packs
kulshan report --packs all --regions us-east-1  # Full 10-pack diagnostic
kulshan history                         # View past scans
kulshan shell                           # Interactive REPL

All 10 Audit Packs

Pack What it detects
cost Cost trends, anomalies, commitment gaps (default)
security IAM misconfigurations, encryption gaps, network exposure
sweep Orphaned and idle resources (waste detection)
dr Backup coverage gaps, multi-AZ gaps, single points of failure
age EOL runtimes, expiring certificates, stale resources
drift CloudFormation drift, IaC coverage gaps
tag Tag compliance violations, unattributed spend
pulse Observability gaps, missing alarms
limit Service quota headroom issues
topo VPC topology issues, CIDR overlaps, route integrity problems
kulshan report --packs cost,security,sweep --regions us-east-1
kulshan report --packs all --regions us-east-1,us-west-2

CUR / Data Export Investigations

Kulshan can investigate cost movements directly from AWS Cost and Usage Report (CUR) or Data Export Parquet files, both locally and from S3. No Athena, no Glue, no data warehouse required. DuckDB queries your data in place.

Validate and Inspect

# Validate CUR Parquet structure
kulshan cur validate --path ./cur/

# Validate S3 manifest
kulshan cur validate --s3 s3://bucket/prefix/ --month 2024-06

# Inspect schema mapping
kulshan cur schema --path ./cur/

# Check S3 readiness (no data download)
kulshan cur s3-check --s3 s3://bucket/prefix/

Investigate Cost Movements

# Local investigation
kulshan investigate cost --path ./cur/ --month 2024-06

# S3 investigation (queries in place via DuckDB httpfs)
kulshan investigate cost --s3 s3://bucket/prefix/ --month 2024-06

# EC2-specific investigation
kulshan investigate ec2 --cur ./cur/ --month 2024-06

# Export as JSON or Markdown
kulshan investigate cost --path ./cur/ --month 2024-06 -o report.json
kulshan investigate ec2 --cur ./cur/ --month 2024-06 -o report.md

What Investigations Include

Cost Investigation: Top movers by service, account, region, usage type. Period-over-period delta with percentages. Suggested next steps for deeper analysis.

EC2 Investigation: Instance family, region, and pricing model breakdowns. Resource-level contributors. Tag coverage analysis.

All investigation outputs include structured evidence with unique IDs, schema version, timestamps, and human_review_required: true.


MCP Server (Agent Integration)

Kulshan exposes its findings to MCP-compatible agents (Claude Desktop, Cursor, Kiro, and others). The agent may reason over that evidence; Kulshan itself returns deterministic, inspectable evidence.

kulshan mcp-serve

MCP Configuration

Add to your MCP client configuration (e.g., .kiro/settings/mcp.json or Claude Desktop config):

{
  "mcpServers": {
    "kulshan": {
      "command": "kulshan",
      "args": ["mcp-serve"]
    }
  }
}

Available MCP Tools

Tool Parameters Description
kulshan_doctor (none) Check AWS caller identity
kulshan_report packs, days, regions Run audit packs and return findings
kulshan_quick_security region Fast security scan of a single region
kulshan_list_packs (none) List available audit packs
kulshan_cur_validate cur_path Validate local CUR Parquet
kulshan_investigate_ec2 cur_path, month Investigate EC2 costs from local CUR
kulshan_investigate_cost s3_uri, month Investigate costs from S3 CUR

Kulshan produces deterministic, inspectable evidence that humans and AI systems can verify.


Output Formats

kulshan report                          # Terminal
kulshan report -o report.html           # HTML (self-contained, shareable)
kulshan report --format json            # JSON
kulshan report --format sarif           # SARIF (GitHub Security tab compatible)
kulshan report --format csv             # CSV (spreadsheet)

Convert a previous scan to a different format without re-running:

kulshan convert -i previous-scan.json --format html -o report.html

Trust and Security

Read-only by design. No write permissions required. Published IAM policy included.

Kulshan is read-only by construction, not read-only by default. There is no cleanup mode to leave off, no write path to enable, and no telemetry to opt out of. Nothing to disable, because nothing exists. The published IAM policy contains zero actions that create, modify, or delete resources. Read every line, verify everything.

  • Every action in the published policy is read-level per the AWS service authorization reference
  • 147 read-only actions, zero write actions
  • Reports stay on your machine, no uploads
  • No telemetry, no phone-home
  • Open source: Apache 2.0

View the IAM policy


AWS API Cost

Typical run cost: approximately $0.15 to $0.25 in AWS Cost Explorer API charges. AWS bills Cost Explorer requests at $0.01 per request. All non-cost packs use free AWS APIs.

The CLI asks for confirmation before running Cost Explorer queries. Use --yes for CI/CD.


Performance

Kulshan uses parallel execution for audit packs and region scanning. A full 10-pack scan completes substantially faster than sequential execution.


About the Name

Kulshan is the Lummi name for the mountain known colonially as Mt. Baker, meaning "great white watcher." We acknowledge the Lummi and Nooksack peoples as the original namers of this mountain.


Built by

Mission FinOps | Mission, BC, Canada

For questions about what Kulshan detects, or for help investigating and explaining findings, contact Mission FinOps.


Links


License

Apache 2.0. Free and open source forever.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kulshan-0.3.0.tar.gz (307.8 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

kulshan-0.3.0-py3-none-any.whl (397.4 kB view details)

Uploaded Python 3

File details

Details for the file kulshan-0.3.0.tar.gz.

File metadata

  • Download URL: kulshan-0.3.0.tar.gz
  • Upload date:
  • Size: 307.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for kulshan-0.3.0.tar.gz
Algorithm Hash digest
SHA256 b4f10a5d7de1432f78f68d59bf00ca29ce2bdc58f30335060491fdd3db683eeb
MD5 ff692df5271dfe149cd6873c432b4227
BLAKE2b-256 961ee1620e575aa8bc42b0c29e236d4bcc70c8ce3c03373c1e665dd3b5e69add

See more details on using hashes here.

Provenance

The following attestation bundles were made for kulshan-0.3.0.tar.gz:

Publisher: publish.yml on azz-kikkr/kulshan

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file kulshan-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: kulshan-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 397.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for kulshan-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 cfe4e38ae7bda01f7b19a3d71d009f92bbd5615637cab28ff684cb24b0d434fd
MD5 408cd0ef5ef26e1c547d4342548258a3
BLAKE2b-256 38d3b13765b6fa3f0eb7284f235248ecb3379ce684cc7a459c6d8784e9bb83cf

See more details on using hashes here.

Provenance

The following attestation bundles were made for kulshan-0.3.0-py3-none-any.whl:

Publisher: publish.yml on azz-kikkr/kulshan

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page