fix(doctor): correct warning count in example after PR #59 icon change

[TechNickAI]

Summary

• Fixes summary line in doctor example output: 2 warnings ⚠️ → 1 warning ⚠️
• PR fix(doctor): correct ℹ️ icon and generalize hook example from PR #56 review #59 changed mcp-debug from ⚠️ to ℹ️, but the summary count wasn't updated to match
• With mcp-debug now informational, only todo-persist.sh is a warning in the example

Bot Feedback Addressed (from PR #59)

Fixed (2):

• cursor[bot] (line 704): "Summary warning count inconsistent after icon change" — ✅ Fixed
• chatgpt-codex-connector[bot] (line 704): "Align summary counts with info-level mcp-debug example" — ✅ Fixed

Declined (1):

• claude[bot] observation: hardcoded todo-persist.sh in example output — 👎 Declined: concrete examples should use concrete filenames; <hook-name>.sh placeholder is correct in the behavioral spec instructions but not in illustrative sample output

🤖 Generated with Claude Code

[claude]

Code Review — PR #63: fix(doctor): correct warning count in example after PR #59 icon change

Overview

This PR makes two types of changes to plugins/core/commands/ai-coding-config.md:

1. Core fix: Corrects the summary count in example output from 2 warnings ⚠️ to 1 warning ⚠️ after PR fix(doctor): correct ℹ️ icon and generalize hook example from PR #56 review #59 demoted mcp-debug from warning to informational.
2. Reformatting: Reflowing long lines to a consistent width and inline-collapsing closing XML-like tags.

---

✅ What's Good

• Correct bug fix: The 2 warnings → 1 warning change is accurate. With mcp-debug now ℹ️, only todo-persist.sh remains as a ⚠️ warning in the example — the count was genuinely wrong.
• Clear PR rationale: The description traces the issue directly back to PR fix(doctor): correct ℹ️ icon and generalize hook example from PR #56 review #59's icon change, making the change easy to audit.
• Declined suggestion is well-reasoned: Keeping todo-persist.sh (a concrete filename) in illustrative sample output instead of a <hook-name>.sh placeholder is the right call — examples should be concrete.

---

⚠️ Minor Issues / Observations

Inline closing tags are an unusual style shift

Several closing tags were moved from their own line to inline with the preceding sentence:

-Report which context was detected at the top of output.
-</context-detection>
+Report which context was detected at the top of output. </context-detection>

Same pattern for </auto-fix> and </output-format>. Since these are LLM-facing instruction blocks (not strict XML), it won't break anything — an LLM parses the closing tag the same either way. But it's an inconsistency with the rest of the file's style where most other closing tags (</checks>, </doctor-mode>, etc.) sit on their own lines. If a formatter is enforcing this, it might be worth checking whether the formatter config is intentional.

Reformatting scope is larger than the fix

The diff includes ~48 lines of pure reflowing (line-wrap changes, added blank lines before list items) unrelated to the count fix. This is low risk since it's a doc-only file, but mixing cosmetic reformatting with a targeted fix makes the diff harder to review and could cause merge conflicts with other concurrent changes. Consider separating formatting-only PRs when practical.

---

No Issues Found

• Correctness: The 1-warning count matches the example output accurately.
• Security: N/A — documentation change only.
• Tests: N/A — no executable code changed.
• Logic: Example output now internally consistent (1 failure ❌, 1 warning ⚠️, 11 passed ✅).

---

Summary

The core fix is correct and necessary. The reformatting is harmless but slightly noisy. Overall this is a straightforward, low-risk documentation correction — good to merge once the inline-tag style choice is confirmed as intentional.

[TechNickAI]

Both observations are accurate. The closing tag inlining and line reflowing were done by the post-edit prettier formatter hook that runs automatically — not intentional style choices. The inconsistency is real and tracked in #64. The targeted fix (1 warning count) is isolated to line 708; the surrounding reformatting is formatter noise. Nothing to change in this PR.