Rich markup silently deletes bracketed path segments (`[id]`) from every printed path

[Wynelson94]

Problem

Paths are printed through rich.console.Console.print, which parses [...] as markup. Any path containing a bracketed segment — every Next.js dynamic route — has that segment silently deleted from the output.

$ python3 -c "from rich.console import Console; Console().print('src/app/api/prospects/[id]/route.ts')"
src/app/api/prospects//route.ts

The data on disk is correct. Only the display lies — and it lies in the worst possible way, because src/app/api/prospects//route.ts collapses on POSIX to src/app/api/prospects/route.ts, a different file that also exists. A reader following fieldnotes get goes and reads the wrong route and concludes the note is wrong.

Where it shows

Every command that prints a reference path: get, stale, verify, brief, touched, for, diff, gaps. Real output from a production Next.js repo:

0001 audit-log-fail-fast              stale:src/app/api/prospects//rou…
0031 rls-with-check-gates-rows-not-co… stale:src/app/api/meetings//rout…

Both notes actually pin [id] routes.

Fix

markup=False (and probably highlight=False) on path output, or rich.markup.escape() on the interpolated path. The pattern is already in the codebase at cli.py:800 for diff bodies:

console.print(rd.diff, markup=False, highlight=False)

Representative sites: cli.py:702, :714, :796, :945, :993. There are ~101 console.print calls total, so a helper (print_path() / escaping at the formatting boundary) beats patching each one.

Corrects the diagnosis in #2

Issue #2's "Related parsing artifact" section attributes the doubled slash to churn_map in gaps.py not normalizing git log --name-only rename entries. That's not it — the artifact appears in commands that never touch git log, and the repro above is pure Rich. That section of #2 can be dropped in favor of this issue.

Found while healing 10 stale notes in a Next.js repo on 2026-07-24.