POLICY WORKBENCH

Your sources.
Your questions.
Visible evidence.

Build a policy library with guided forms and plain text files. See what needs attention, inspect exact source lines and prepare a portable review packet.

Works on this device. No account connection. Save your work before closing.

Loading the local workbench…

01 · REVIEW SCOPE

Name this source collection

Fictional input library

Size is calculated from the complete current input library.

SOURCE DETAILS

Selected source

Source declarations supplied by you

New sources start with all three off. These statements are not independently verified. Any blocking issue suppresses evidence across the collection.

Plain UTF-8 .txt or .md only. No PDF, DOCX, OCR or URL retrieval. Maximum 120,000 characters, 5,000 lines and 2,000 characters per line per source.

03 · LITERAL EVIDENCE REQUESTS

Questions and exact search phrases

Describe what a human reviewer needs to examine, then enter one keyword or phrase per line. The engine searches those exact phrases within individual source lines. It does not infer meaning, synonyms or an answer.

04 · CURRENT REVIEW

Readiness, evidence and follow-ups

Review needed

Review the current library to see its source readiness and complete evidence report. Any input or handoff declaration edit clears results and disables report downloads.

05 · PORTABLE REVIEW PACKET

Keep the complete handoff

Every result download checks the current engine before it is prepared. The Review Desk draft is available when this library fits its 256 KiB input bound. Saved input libraries remain separate from reports.

Prepare a current review to check the complete Review Desk draft size.

A collection you can inspect

Only files you explicitly select are read. Imported text stays inert. This page does not connect accounts, send source text to a model, upload data or persist browser storage. Save the input library to keep editing; save a report to give a human reviewer the complete evidence packet.