Existing dependency analysis approaches primarily use metric-driven assessment, vulnerability detection, visualization dashboards, or prediction models. However, they often lack interpretability in explaining why certain dependencies warrant further attention. To address this limitation, we propose Investigative Dependency Reasoning Framework (IDRF), a prototype system designed to analyze python package dependencies using structural metrics and metadata collected from the python package index (PyPI) and related repositories. IDRF evaluates twelve bounded dependency environments comprising 318 packages, constructs bounded dependency graphs with limited-depth expansion, and computes investigative priority scores using heuristic metrics. The system produces human-readable investigative insight cards highlighting dependencies requiring further analysis based on interpretable scoring rules, generating 60 insight cards across the evaluated environments. Results show that IDRF can surface dependencies that warrant additional attention within bounded dependency constraints. It demonstrates how bounded graph analysis coupled with priority scoring produces interpretable dependency reasoning.