# Independent reproduction protocol WSA-20260831

**Published:** 2026-08-31  
**Target run:** `ADR-WSA-20260825`  
**Status:** Open protocol; no independent reproduction received as of 2026-08-31

## Purpose

This protocol lets an unaffiliated analyst test the published sentence, trail, and stop checks without accepting InfiniSynapse’s interpretation. It is not itself an independent report.

## Materials

- Published desk log `ADR-WSA-20260825`
- Aggregate CSV `aggregate-ADR-WSA-20260825.csv`
- Verification script `verify-ADR-WSA-20260825.py`
- A separately obtained authorized dataset or documented public extract

Do not request or publish the private sanitized composite behind the first-party desk run. Use public data or data you are authorized to inspect.

## Procedure

1. Declare identity, affiliation, conflicts, retrieval date, source version, and file hashes.
2. Predeclare the decision, grain, clock, filters, exclusions, stop condition, and expected output columns.
3. Run a chat-only control and score: sentence written, trail opened, stop named.
4. Run a sentence-plus-trail pass and retain the source slice, filter list, calculation, output, and stop decision.
5. Have a second reviewer recalculate the decision-driving subtotal without reading the narrative conclusion.
6. Publish methods, results, failed runs, limitations, and enough non-sensitive artifacts for inspection.

## Acceptance criteria

An independent report must be authored outside InfiniSynapse, disclose conflicts, identify the source version, reproduce its own arithmetic, and state whether its three artifact checks agree with or contradict the page. Merely running the first-party CSV verification script does not count.

Send a public report URL or a documented contradiction to [zhuhl@infinisynapse.com](mailto:zhuhl@infinisynapse.com). InfiniSynapse may link qualifying reports whether they confirm or contradict the first-party run.