AWS IAM Exposure Review is a local pipeline with explicit boundaries:
AWS read-only APIs
-> normalized inventory
-> effective permission evaluation
-> typed attack graph
-> research-attributed path rules
-> stable findings and lifecycle comparison
-> offline HTML/JSON/CSV/SARIF evidence
collector: gathers IAM authorization, credential metadata, trust, Access Analyzer, instance profile, organization, and SCP context where permitted.policy: normalizes statements and evaluates allow, deny, boundary, SCP, session, resource, condition, and target uncertainty.graph: creates typed nodes and relationships for principals, policy sources, grants, actions, resources, paths, and gained capabilities.rules: evaluates versioned attack-path rules and emits explainable evidence.lifecycle: creates stable path fingerprints and compares complete scans.report: writes self-contained assessment artifacts without external assets or tracking.updates: validates rule schemas and explicit SHA-256-pinned updates.
The supported Python API is exported from iam_exposure.api and the package root. Dictionary outputs are versioned by schema_version; fields may be added in a minor release, while removals require a major release.
Hosted products must consume a released package version. The open-source package does not import hosted-service code, enforce commercial licenses, or contact a production service.