fix(ego-lint): avoid false FAIL when enum-id precedes schema-id in check-schema (#148)

## Problem

`check-schema.sh` extracts the schema ID using:
```bash
grep -oP 'id="[^"]*"' "$schema_file" | head -1 | sed 's/id="//;s/"//'
```
This grabs the **first** `id=` attribute in the file — which may belong
to a `<enum>` element that appears before the `<schema>` element.

**Reproduced on caffeine** (`caffeine@patapon.info`): the schema file
defines three `<enum id="...">` elements before `<schema
id="org.gnome.shell.extensions.caffeine">`. The old code extracted the
enum ID (`org.gnome.shell.extensions.caffeine.context-control`), causing
two false FAILs:
- `schema/id-matches` — ID didn't match `metadata.json settings-schema`
- `schema/filename-convention` — expected filename based on wrong ID

## Fix

Introduce `extract_schema_id()` helper that pre-filters to lines
matching `<schema\b` before extracting `id=`:
```bash
extract_schema_id() {
    local file="$1"
    grep -P '<schema\b' "$file" | grep -oP 'id="[^"]*"' | head -1 | sed 's/id="//;s/"//'
}
```
The old inline one-liner is replaced at all three call sites
(`schema/id-matches`, `schema/filename-convention`,
`schema/gnome-trademark`).

## Test

Before fix on caffeine:
```
FAIL|schema/id-matches|Schema ID 'org.gnome.shell.extensions.caffeine.context-control' does not match ...
FAIL|schema/filename-convention|Schema filename 'org.gnome.shell.extensions.caffeine.gschema.xml' MUST be '...'
```

After fix on caffeine:
```
PASS|schema/id-matches|Schema ID 'org.gnome.shell.extensions.caffeine' matches metadata.json settings-schema
PASS|schema/filename-convention|Schema filename matches ID: org.gnome.shell.extensions.caffeine.gschema.xml
```

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Solomon <sol@nanoclaw.dev>

Author: ZviBaratz

skills/ego-lint/scripts/check-schema.sh

@@ -8,6 +8,16 @@
 set -euo pipefail
 
 EXT_DIR="$(cd "${1:-.}" && pwd)"
+
+# Extract the schema id from a .gschema.xml file.
+# Uses a <schema>-specific grep to avoid picking up <enum id=> or other
+# elements that appear before <schema id=> in files with multiple top-level
+# elements (e.g. caffeine's schema has three <enum> elements first).
+extract_schema_id() {
+    local file="$1"
+    grep -P '<schema\b' "$file" | grep -oP 'id="[^"]*"' | head -1 | sed 's/id="//;s/"//'
+}
+
 METADATA="$EXT_DIR/metadata.json"
 # src/ layout fallback
 [[ ! -f "$METADATA" && -f "$EXT_DIR/src/metadata.json" ]] && METADATA="$EXT_DIR/src/metadata.json"
@@ -53,8 +63,8 @@ echo "PASS|schema/exists|Found ${#schema_files[@]} schema file(s)"
 # Validate schema IDs match metadata
 if [[ "$has_settings_schema" == true ]]; then
     for schema_file in "${schema_files[@]}"; do
-        # Extract schema id from XML
-        schema_id="$(grep -oP 'id="[^"]*"' "$schema_file" | head -1 | sed 's/id="//;s/"//')"
+        # Extract schema id from XML (use extract_schema_id to avoid matching <enum id=> etc.)
+        schema_id="$(extract_schema_id "$schema_file")"
         if [[ "$schema_id" == "$settings_schema" ]]; then
             echo "PASS|schema/id-matches|Schema ID '$schema_id' matches metadata.json settings-schema"
         else
@@ -65,7 +75,7 @@ fi
 
 # Validate schema filename convention: <schema-id>.gschema.xml
 for schema_file in "${schema_files[@]}"; do
-    schema_id="$(grep -oP 'id="[^"]*"' "$schema_file" | head -1 | sed 's/id="//;s/"//')"
+    schema_id="$(extract_schema_id "$schema_file")"
     if [[ -n "$schema_id" ]]; then
         expected_filename="${schema_id}.gschema.xml"
         actual_filename="$(basename "$schema_file")"
@@ -97,7 +107,7 @@ done
 
 # Check for GNOME trademark in schema IDs
 for schema_file in "${schema_files[@]}"; do
-    schema_id="$(grep -oP 'id="[^"]*"' "$schema_file" | head -1 | sed 's/id="//;s/"//')"
+    schema_id="$(extract_schema_id "$schema_file")"
     if [[ -n "$schema_id" ]]; then
         # Strip the standard prefix (case-insensitive), then check for 'gnome' in the extension-specific part
         lower_id="${schema_id,,}"