fix: cross-section search and CSS loading for chat widget

- agent.go: add section="search" that queries experience, projects,
  skills, and courses simultaneously — fixes missing results when
  a technology spans multiple CV sections (e.g. Java at Insa)
- head-styles.html: use modular CSS in development mode and load
  chat CSS separately — fixes unstyled page when bundle is stale

internal/chat/agent.go

@@ -33,6 +33,7 @@ You answer questions about the CV owner's experience, projects, skills, educatio
 
 RULES:
 - Use the query_cv tool to look up CV data before answering. Never make up information.
+- For technology questions (e.g. "Java", "Go", "React"), ALWAYS use section="search" — this searches across experience, projects, courses, and skills simultaneously. Do NOT search only projects or only experience.
 - Answer in the SAME LANGUAGE the user writes in. If they ask in Spanish, answer in Spanish.
 - Be concise and direct — visitors want quick answers, not essays.
 - When listing items (projects, technologies, companies), use bullet points.
@@ -42,18 +43,19 @@ RULES:
 - You represent the CV owner professionally — be friendly but not overly casual.
 
 EXAMPLES of questions you might receive:
-- "How many years of experience does Juan have?"
-- "What Go projects has he built?"
-- "Has he worked with React?"
-- "Tell me about his time at Olympic Broadcasting"
-- "What certifications does he have?"`,
+- "How many years of experience does Juan have?" → section="summary"
+- "What Java experience does he have?" → section="search", query="java"
+- "Has he worked with React?" → section="search", query="react"
+- "Tell me about his time at Olympic Broadcasting" → section="search", query="olympic"
+- "What certifications does he have?" → section="certifications"
+- "List all his projects" → section="projects"`,
 		Tools: []tool.Tool{queryTool},
 	})
 }
 
 // QueryCVArgs is the input for the CV query tool.
 type QueryCVArgs struct {
-	Section  string `json:"section" jsonschema:"CV section to query: 'experience', 'projects', 'skills', 'education', 'languages', 'certifications', 'courses', 'awards', 'summary', 'all'"`
+	Section  string `json:"section" jsonschema:"CV section to query: 'search' (cross-section keyword search — recommended for technology queries), 'experience', 'projects', 'skills', 'education', 'languages', 'certifications', 'courses', 'awards', 'summary', 'all'"`
 	Query    string `json:"query" jsonschema:"Search term to filter results (e.g. 'Go', 'React', '2019', 'Olympic'). Empty returns all items in the section."`
 	Language string `json:"language" jsonschema:"Language for CV data: 'en' or 'es'. Default: 'en'."`
 }
@@ -71,6 +73,7 @@ func newQueryCVTool(dataCache *cache.DataCache) (tool.Tool, error) {
 		Name: "query_cv",
 		Description: `Query the CV data to answer questions about experience, projects, skills, education, certifications, and more.
 Use the 'section' parameter to target a specific area, and 'query' to filter by keyword.
+For technology or keyword queries (e.g. "Java", "Go", "React", "Olympic"), use section="search" to search across experience, projects, skills, and courses simultaneously. This avoids missing results that appear in multiple sections.
 Always call this tool before answering CV-related questions.`,
 	}, func(ctx tool.Context, args QueryCVArgs) (QueryCVResult, error) {
 		lang := args.Language
@@ -128,6 +131,31 @@ Always call this tool before answering CV-related questions.`,
 			result.TotalFound = len(cv.Awards)
 			result.Data = mustJSON(cv.Awards)
 
+		case "search":
+			// Cross-section search: search across experience, projects, skills, and courses simultaneously.
+			crossResult := make(map[string]any)
+			total := 0
+
+			if expMatches := filterExperience(cv.Experience, q); len(expMatches) > 0 {
+				crossResult["experience"] = expMatches
+				total += len(expMatches)
+			}
+			if projMatches := filterProjects(cv.Projects, q); len(projMatches) > 0 {
+				crossResult["projects"] = projMatches
+				total += len(projMatches)
+			}
+			if skillMatches := filterSkills(cv.Skills, q); len(skillMatches) > 0 {
+				crossResult["skills"] = skillMatches
+				total += len(skillMatches)
+			}
+			if courseMatches := filterCourses(cv.Courses, q); len(courseMatches) > 0 {
+				crossResult["courses"] = courseMatches
+				total += len(courseMatches)
+			}
+
+			result.TotalFound = total
+			result.Data = mustJSON(crossResult)
+
 		case "all":
 			// Return a high-level overview
 			overview := map[string]int{