Problem

workflow_resolve in src/specify_cli/workflows/overlays/_commands.py prints each layer like this:

f"  • [{layer.tier}] {layer.source} (priority={priority})"

Rich parses [base] and [project-overlay] as style tags, so the tier column is silently swallowed — on every single run, not just for odd input.

Reproduction on current main (81bf741)

BEFORE: '*  .specify/workflows/wf/workflow.yml (priority=n/a)'
AFTER : '* [base] .specify/workflows/wf/workflow.yml (priority=n/a)'

BEFORE: '*  ov1.yml (priority=10)'
AFTER : '* [project-overlay] ov1.yml (priority=10)'

Note the double space in the BEFORE lines — that is where the tier used to be. The layer's tier is one of the two things workflow resolve exists to tell you (which layer, at what precedence), so the human-readable output is missing half its answer. The --json payload carries tier correctly; only the printed form loses it.

Why the existing test didn't catch it

test_workflow_resolve asserts:

assert "base" in result.output

That passes with the bug, because the step-attribution section prints the same word as a step source. The tier has to be asserted in bracket form, on the layers line itself — which is what the new test does.

Fix

Escape the literal bracket with \[, exactly as presets/_commands.py:378 already does:

# presets/_commands.py:374-378
# Escape the literal bracket (\[) so Rich renders `[<strategy>]`
f"    {i + 1}. \[{_escape_markup(str(strategy_label))}] "

I also escaped the interpolated values in the same output block — layer.source is a filesystem path and composed.step_id / composed.source are user-authored ids, all of which can contain brackets and would break the same lines. That keeps the whole of this one command's output correct rather than fixing half of it.

No breaking change. rich.markup.escape is a no-op on bracket-free values, and the --json payload is untouched. The only difference is that output which previously lost characters now shows them.

Verification

• Fail-before / pass-after: the new test fails on unpatched src and passes with the fix. File: 1 failed → 21 passed.
• Scoped regression over tests/workflows: failure set identical to the clean-main baseline captured on 81bf741 (10 pre-existing in scope, all Windows symlink-privilege).
• uvx ruff@0.15.0 check src tests → clean

Written with assistance from Claude Code. Bug found, reproduced, and verified by me on current main.