# Handwritten statements to text kit ## What this pack lets workers do A collection of nouns and verbs for turning handwritten statements and forms dropped into a folder into text typed exactly as written, keeping the writer's slang, spelling, acronyms and words, flagging every uncertain reading for a person, and listing acronyms against a local glossary, while staff check every flag, follow up on safety, sign off and decide any use. Type each handwritten page in a drop folder exactly as written, keeping slang, spelling, grammar, acronyms, crossings-out and the writer's own words, and flag every word the agent is unsure of. A local acronym glossary helps readers without ever changing the text. A person checks every flag against the page, trained staff follow up on anything about safety, and the program lead signs off and decides how a statement is used. ## Nouns | Noun | What it holds | | --- | --- | | Kit | The agent's whole setup as one noun: the step map and pins to every part below | | Skill `verbatim-transcriber` | Types handwritten pages exactly as written and flags every uncertain reading for a person. | | Skill `acronym-glossary-assistant` | Lists acronyms and local terms in a transcription with their glossary meanings, without changing the text. | | Read-only tool card `list_dropped_pages` | Lists the page images in the intake drop folder with page numbers, sizes and SHA-256 digests, without moving or changing them. | | File: Verbatim Transcription Rules | The rules every transcription follows: type exactly what is written, flag what is uncertain, and leave the writer's words alone. | | File: Transcription Sheet | One row per written line: the batch, page, line, the text as written, whether it holds a flag, the alternatives and whether a reviewer checked it. | | File: Local Acronym Glossary | Acronyms, local terms and common writing abbreviations with their meanings, as a reading aid that never changes a transcription. | | Checklist | 5 checks a person runs with the kit | | Sources | 1 dated public sources | | Test cases | 4 synthetic cases; 2 must go to a person | | Need, in its own words | One or two plain sentences an organization might say about this work, marked synthetic | | Templates | one page typed exactly as written, with every flag located, acronyms and local terms found in a transcription, and a person's step record | ## Verbs | Verb | Who | What happens | | --- | --- | --- | | `transcribe-pages` | agent | Type each page exactly as written | | `list-acronyms` | agent, no model | List acronyms and local terms against the glossary | | `verify-readings` | person | Check every flagged reading against the original page | | `safety-follow-up` | person | Follow up in person when a statement raises safety | | `sign-off-statement` | person | Sign off the transcription and decide how it is used | | `share-statement` | person | Share a statement outside the organization | Each verb can also end with `no_change`, `unknown` or `blocked`, and can propose the next verb. Verbs for a person are claimed by people, never by an agent. ## How to use it Write the nouns in [nouns.json](nouns.json) create-only with a key scoped to your team's surface, then fill a template from [verbs.json](verbs.json) with `scripts/build_packs.py --instantiate