docs: quote SKILL.md YAML description and use $TOKEN in header examples

Author: firede

README.md

@@ -133,7 +133,7 @@ agent-fetch --mode static --meta=false https://example.com
 agent-fetch --mode raw https://example.com
 
 # Authenticated request
-agent-fetch --header 'Authorization: Bearer <token>' https://example.com
+agent-fetch --header "Authorization: Bearer $TOKEN" https://example.com
 
 # Batch fetch with concurrency control
 agent-fetch --concurrency 4 https://example.com https://example.org

README.zh.md

@@ -133,7 +133,7 @@ agent-fetch --mode static --meta=false https://example.com
 agent-fetch --mode raw https://example.com
 
 # 带认证请求
-agent-fetch --header 'Authorization: Bearer <token>' https://example.com
+agent-fetch --header "Authorization: Bearer $TOKEN" https://example.com
 
 # 批量抓取，控制并发
 agent-fetch --concurrency 4 https://example.com https://example.org

skills/agent-fetch/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agent-fetch
-description: Fetch web pages as clean Markdown via the agent-fetch CLI. Handles cases that built-in web-fetch tools may not cover well: JavaScript-rendered pages (headless browser), authenticated endpoints (custom headers), batch fetching multiple URLs concurrently, and content output without AI summarization. Also works standalone when no built-in fetch tool is available.
+description: "Fetch web pages as clean Markdown via the agent-fetch CLI. Handles cases that built-in web-fetch tools may not cover well: JavaScript-rendered pages (headless browser), authenticated endpoints (custom headers), batch fetching multiple URLs concurrently, and content output without AI summarization. Also works standalone when no built-in fetch tool is available."
 ---
 
 # Agent Fetch