A Software-First Open-Source Educational Instrumentation Workflow for Reproducible Wireless Packet Observability

Reproducible packet-level wireless experimentation requires more than access to a testbed or a traffic generator: the configuration, packet evidence, metric semantics, software state, and provenance of an execution must remain connected. Existing physical testbeds, emulators, traffic generators, and general reproducibility frameworks typically address these concerns at different layers, leaving a practical gap for a lightweight workflow that couples packet-level evidence with explicit observation-level semantics. This paper presents a software-first open-source workflow in which a declarative profile is converted into a canonical packet table, analyzed separately at per-node and merged-stream levels, and accompanied by manifests and integrity hashes. The workflow is evaluated through fixed-input determinism, configuration-to-record consistency, metric checks, 30-seed stability, controlled-omission and offered-load sensitivity, a local reference scalability benchmark, and automated fresh-environment continuous-integration executions. For the tagged artifact used by this study, the CI workflow successfully installs the software, runs the test suite, executes a synthetic demonstration, and verifies deterministic reference traces on Ubuntu with Python 3.9, 3.11, and 3.12. With omissions disabled, generated record counts match the configured event model and the packet traces contain no missing or duplicate sequence identifiers. The revised span-based event-rate estimator uses the N−1 observed inter-event intervals rather than N records, preventing the systematic finite-span inflation of the previous formulation. Across 30 seeds per profile, span-rate coefficients of variation remain below 0.007%; controlled omission tests expose the expected boundary limitation of sequence-span completeness; and a local reference benchmark scales from 6000 to 120,000 records with approximately linear generation, analysis, and CSV-storage growth. The aggregation-ratio calculation is treated as an internal consistency invariant rather than as independent validation. The resulting workflow is intended to help researchers inspect the configuration-to-evidence chain and to help educators teach provenance and aggregation-aware interpretation without conflating synthetic traces with PHY/MAC or radio measurements. The present evaluation remains synthetic and does not establish physical-link fidelity, capture accuracy, deployment performance, or learning gains.