adhd-filterlisted

## Purpose Take any notes, document, or pasted conversation and restructure it for ADHD-friendly reading without losing a single fact, equation, nuance, or definition. The information is identical — only the presentation layer changes. **Non-negotiable:** lossless information quality. No paraphrasing. No removal. No invented content. --- ## Argument Resolution `$ARGUMENTS` may be: | Input | Behavior | |---|---| | A file path (e.g. `notes/calculus.md`) | Process that exact file | | A glob (e.g. `notes/*.md`) | Process every matching file not already filtered | | A directory (e.g. `notes/`) | Process every `.md` inside it not already filtered | | Empty, with text pasted in the conversation | Format the pasted text and return it inline (no file written) | A file counts as "already filtered" if its frontmatter contains `adhd_filtered: true`. Pass `--force` to reprocess anyway. If no match is found, report: "No files matched — check the path or glob." and stop. --- ## Procedure ### Step 1 — Identify target files Based on argument resolution above, collect the list of files to process. For a glob or directory, skip any file whose frontmatter already has `adhd_filtered: true` (unless `--force` is passed). List the targets to the user: "Filtering N file(s): [names]" If the input is pasted text instead of a path, skip file handling — run Steps 3–5 on the text and return the result inline. --- ### Step 2 — Read each target file Read each target file. For each, perform