Fix token estimation: include ToolUse arguments in text_length

`MessageContent::text_length()` returned 0 for `ToolUse` blocks,
ignoring the tool name and JSON input arguments. This caused the
compactor's `estimate_token_count()` (which uses `text_length()`)
to massively undercount tokens when conversations contained tool
calls with large arguments (e.g. web_search results, page content).

The result: compaction never triggered despite the session exceeding
the context window, leading to "Token limit exceeded" errors.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

crates/openfang-types/src/message.rs

@@ -142,8 +142,10 @@ impl MessageContent {
                     ContentBlock::Text { text, .. } => text.len(),
                     ContentBlock::ToolResult { content, .. } => content.len(),
                     ContentBlock::Thinking { thinking } => thinking.len(),
-                    ContentBlock::ToolUse { .. }
-                    | ContentBlock::Image { .. }
+                    ContentBlock::ToolUse { name, input, .. } => {
+                        name.len() + input.to_string().len()
+                    }
+                    ContentBlock::Image { .. }
                     | ContentBlock::Unknown => 0,
                 })
                 .sum(),