fix(excalidraw-diagram-generator): force UTF-8 stdout so add-arrow/add-icon don't crash on Windows

[brainit-consulting]

Problem

On Windows, add-arrow.py and add-icon-to-diagram.py crash at the very end of an otherwise
successful run. Both finish by printing a check mark, e.g.:

print(f"✓ Successfully added arrow to diagram")

On a standard Windows console Python's stdout uses the cp1252 code page, which cannot encode
U+2713 (✓). That print raises UnicodeEncodeError, so the script exits with code 1 even though it
has already saved the file. The edit lands, but any caller that checks the exit code sees a failure.
It is also easy to misdiagnose, because the traceback goes to stdout while stderr stays empty.

Repro (Windows, default code page):

python add-arrow.py diagram.excalidraw 0 0 100 100 --label "x"
...
Saving diagram
Error: 'charmap' codec can't encode character '✓' in position 0: character maps to <undefined>
# exit code 1, but diagram.excalidraw WAS modified

This makes the two scripts unreliable when driven from automation (or any tool that checks exit codes)
on Windows.

Fix

Reconfigure stdout to UTF-8 at the start of main() in both scripts. This is the same guard already
used by split-excalidraw-library.py in the same scripts/ directory, so this just brings the other
two scripts in line:

if hasattr(sys.stdout, "reconfigure"):
    # Ensure consistent UTF-8 output on Windows consoles.
    sys.stdout.reconfigure(encoding="utf-8")

No behavior change on macOS/Linux; on Windows the success message now prints and the script exits 0.

Verification

On Windows with the default cp1252 console:

• before: add-arrow.py ... -> exit 1, charmap error raised after the file was already saved
• after: add-arrow.py ... -> prints ✓ Successfully added arrow to diagram, exit 0

Two files, +3/-0 each, no functional change beyond the encoding fix.

🤖 Generated with Claude Code

[github-actions]

⚠️ This PR targets main, but PRs should target staged.

The main branch is auto-published from staged and should not receive direct PRs.
Please close this PR and re-open it against the staged branch.

You can change the base branch using the Edit button at the top of this PR,
or run: gh pr edit 1914 --base staged

[github-actions]

🔍 Skill Validator Results

⚠️ Warnings or advisories found

Scope	Checked
Skills	1
Agents	0
Total	1
Severity	Count
---	---:
❌ Errors	0
⚠️ Warnings	1
ℹ️ Advisories	0

Summary

Level	Finding
ℹ️	Found 1 skill(s)
ℹ️	[excalidraw-diagram-generator] 📊 excalidraw-diagram-generator: 5,862 BPE tokens [chars/4: 5,926] (comprehensive ✗), 43 sections, 22 code blocks
ℹ️	[excalidraw-diagram-generator] ⚠ Skill is 5,862 BPE tokens (chars/4 estimate: 5,926) — "comprehensive" skills hurt performance by 2.9pp on average. Consider splitting into 2–3 focused skills.
ℹ️	✅ All checks passed (1 skill(s))

Full validator output

```text Found 1 skill(s) [excalidraw-diagram-generator] 📊 excalidraw-diagram-generator: 5,862 BPE tokens [chars/4: 5,926] (comprehensive ✗), 43 sections, 22 code blocks [excalidraw-diagram-generator] ⚠ Skill is 5,862 BPE tokens (chars/4 estimate: 5,926) — "comprehensive" skills hurt performance by 2.9pp on average. Consider splitting into 2–3 focused skills. ✅ All checks passed (1 skill(s)) ```

[brainit-consulting]

Thanks for the heads-up. Retargeted this PR to staged and rebased the branch onto it, so the diff is just the intended change: the UTF-8 stdout guard in add-arrow.py and add-icon-to-diagram.py (+3/-0 each), matching the guard split-excalidraw-library.py already uses.