STR Execution and Reporting Guide

Status: Final draft prepared for handoff (pending review)
Version: 0.9
Owner: BionicLoop engineering
Prepared by: BionicLoop engineering Reviewer: ____
Approver: ____
Decision date: ____
Effective date: ____
Baseline freeze SHA: ____
Last updated: 2026-04-06 15:20 EDT

Revision History

Version	Date	Author	Summary of Changes
0.1	2026-03-27	Engineering	Initial STR execution guide draft
0.9	2026-04-06	BionicLoop engineering	Added handoff-ready document-control metadata for the software package

This guide defines how protocol execution should be supervised, how evidence should be collected, and how STR-* reports should be authored for human review.

Goal

Every formal verification run should produce a reviewable package that lets a human reviewer answer:

what was executed
on what baseline
in what environment
what evidence was collected
whether the run passed, failed, or deviated

Required Roles Per Formal Run

Protocol owner
owns the STP-* document content and scope
Executor
performs the run
Evidence collector
ensures screenshots, logs, artifacts, and environment metadata are preserved
STR author
assembles the report and ties evidence to TV-* IDs
Reviewer
reviews the STR against the underlying evidence
Approver
signs off the run for submission closure when appropriate

In small runs, one person may hold multiple roles, but the role names should still be recorded.

Required STR Package Contents

Every formal STR-* package should include:

run label and timestamp
git SHA / branch / baseline identifier
executor name
reviewer name
device/simulator/toolchain versions
exact protocol ID (STP-*)
exact verification IDs covered (TV-*)
executed commands or operator steps
expected result summary
actual result summary
pass/fail/deviation decision
linked raw evidence

Raw Evidence Requirements

Minimum evidence types by modality:

Automated / Simulator

command lines
console output
xcresult path
any generated logs

Simulation

run context
expected output
actual output
diff output
results summary

Hardware / Alert Drill

device model and OS version
app build and git SHA
timestamped operator notes
screenshots / screen recordings
local notification evidence where applicable
console/device logs where available
if cloud-log review support was used:
test_run_id
UTC time window
upload threshold used
explicit note that the session was started before execution and stopped after execution

Integration Log Session Rule

When a run depends on CloudWatch log review, the executor must treat the Home Settings Integration Log Session as part of evidence capture, not optional debugging:

start the session before the test scenario begins
stop the session immediately after the scenario ends
record the displayed test_run_id and UTC window in the STR package
do not assume the run ID can be rediscovered from CloudWatch alone until end-to-end metadata preservation is confirmed

STR Authoring Format

Each STR-* README or report should include these sections:

Purpose
Protocol reference
Scope / covered TV-*
Environment
Execution summary
Evidence list
Results by TV-*
Deviations / limitations
Reviewer conclusion

Execution Governance Rules

Do not mark a TV-* row complete without a linked STR path.
Do not substitute working-lane evidence for formal closure evidence.
Do not collapse multiple materially different environments into one report without identifying each environment explicitly.
If a run deviates from the protocol, capture the deviation in the STR instead of silently editing the result narrative.

Current Authoring Intent

BionicLoop engineering will supervise protocol execution and STR authoring so that:

evidence is captured consistently
protocol gaps are discovered before formal review
reviewers receive a package that can be checked against underlying artifacts rather than narrative alone