Remove final_output tool, improve scout agent

- Remove final_output tool to allow LLM responses to stream naturally
- Update system prompts to request summaries instead of tool calls
- Rename final_output_summary to summary in session continuation
- Update tool count tests (12→11 core tools, 27→26 total)
- Delete obsolete final_output tests

Scout agent improvements:
- Simplify WebDriver usage instructions
- Prefer DuckDuckGo/Brave/Bing over Google
- Support passing task directly to agent mode
- Suppress completion message for scout (needs clean output for research tool)

agents/scout.md

@@ -103,13 +103,9 @@ You have access to WebDriver browser automation tools for web research.
 **How to use WebDriver:**
 1. Call `webdriver_start` to begin a browser session
 2. Use `webdriver_navigate` to go to URLs (search engines, documentation sites, etc.)
-3. **IMPORTANT**: Always use `webdriver_get_page_source` with `save_to_file` parameter to save the page HTML to disk
+3. Use all the standard webdriver navigation tools to scan and explore websites 
-4. Read the saved HTML file with `read_file` to extract the information you need
+3. Call `webdriver_quit` when done
-5. Call `webdriver_quit` when done
 
 **Best practices:**
-- Do NOT use `webdriver_screenshot` or try to decode page content visually - always save HTML to disk and read it
+- Do NOT use Google, prefer DuckDuckGo, Brave Search or Bing in that order.
-- Save pages to the `tmp/` subdirectory (e.g., `tmp/search_results.html`)
+- If you're struggling to navigate sites, try saving pages to the `tmp/` subdirectory (e.g., `tmp/search_results.html`), then parse the HTML to find what you need
-- Parse the HTML text content to find what you need
-- For search engines, look for result links and titles in the HTML
-- Close the WebDriver session when you're done to free resources

crates/g3-cli/src/lib.rs

@@ -433,6 +433,7 @@ pub async fn run() -> Result<()> {
             cli.config.as_deref(),
             cli.quiet,
             cli.new_session,
+            cli.task.clone(),
         )
         .await;
     }
@@ -445,6 +446,7 @@ pub async fn run() -> Result<()> {
             cli.config.as_deref(),
             cli.quiet,
             cli.new_session,
+            cli.task.clone(),
         )
         .await;
     }
@@ -653,6 +655,7 @@ async fn run_agent_mode(
     config_path: Option<&str>,
     _quiet: bool,
     new_session: bool,
+    task: Option<String>,
 ) -> Result<()> {
     use g3_core::get_agent_system_prompt;
     use g3_core::find_incomplete_agent_session;
@@ -821,10 +824,16 @@ async fn run_agent_mode(
         // Fresh start - the agent prompt should contain instructions to start working immediately
         "Begin your analysis and work on the current project. Follow your mission and workflow as specified in your instructions."
     };
+    // Use provided task if available, otherwise use the default initial_task
+    let final_task = task.as_deref().unwrap_or(initial_task);
     
-    let _result = agent.execute_task(initial_task, None, true).await?;
+    let _result = agent.execute_task(final_task, None, true).await?;
     
-    output.print("\n✅ Agent mode completed");
+    // Don't print completion message for scout agent - it needs the last line
+    // to be the report file path for the research tool to read
+    if agent_name != "scout" {
+        output.print("\n✅ Agent mode completed");
+    }
     Ok(())
 }