test(ai): expand unit and functional coverage for /ai feature

Unit tests (tests/unittests/test_ai_client.c): +50 cases on top of the
existing 50 — total 100. Covers:
  - chat response parser (ai_parse_response): OpenAI content + Perplexity
    text formats, escape decoding, empty/null/missing inputs, format-string
    safety, multiline content
  - error envelope parser (ai_parse_error_message): standard envelope,
    nested escapes, missing fields, null/empty
  - extended JSON escape: \b, \f, \r, all-specials, UTF-8 pass-through
  - provider autocomplete cycling with >=2 matches and wrap-around
  - session edge cases: NULL args, 100-message order preservation,
    set_model(NULL), ref/unref(NULL)
  - provider edge cases: get/remove with NULL, double-remove, survival
    via session ref after ai_remove_provider
  - settings: multi-key independence, missing key, cross-provider
    isolation
  - model parsing edges: data not array, empty data, "id" outside data,
    multiple models
  - prefs round-trip: set token -> shutdown -> init -> token reloaded
    from disk (uses load_preferences fixture)

Functional tests:
  - Console only (test_ai.c): 15 cases for the /ai command surface that
    don't need HTTP. Covers /ai help, providers list, set provider/token,
    start with/without key/unknown provider, clear, remove, default
    provider/model, switch without window, bad subcommand.

Infrastructure:
  - TEST_GROUPS bumped to 5 in proftest.c; AI tests live in their own
    Group 5 because mixing them with stabber-driven tests in Group 4
    poisoned stbbr_stop() teardown.
  - PROF_FUNC_TEST_AI macro in functionaltests.c registers AI tests
    with ai_init_test() which wraps init_prof_test with a prof_connect()
    so stabber sees a graceful disconnect at teardown.

Source-level changes to enable testing:
  - src/ai/ai_client.{c,h}: dropped 'static' from _parse_ai_response
    (renamed ai_parse_response) and _parse_error_response (renamed
    ai_parse_error_message); declared under a "Parsing helpers (exposed
    for testing)" section. Same approach as ai_parse_models_from_json.

Results in Docker (cproof-debian image):
  - 595/595 unit tests pass
  - 15/15 AI functional tests pass (Group 5)

src/ai/ai_client.c

@@ -1250,8 +1250,8 @@ _build_json_payload_from_list(const gchar* model, GList* history, const gchar* p
     return json_payload;
 }
 
-static gchar*
-_parse_ai_response(const gchar* response_json)
+gchar*
+ai_parse_response(const gchar* response_json)
 {
     if (!response_json || strlen(response_json) == 0)
         return NULL;
@@ -1337,8 +1337,8 @@ _parse_ai_response(const gchar* response_json)
  * Extract error message from an API error JSON response.
  * Handles format: {"error":{"message":"...","type":"...","code":...}}
  */
-static gchar*
-_parse_error_response(const gchar* error_json)
+gchar*
+ai_parse_error_message(const gchar* error_json)
 {
     if (!error_json || strlen(error_json) == 0)
         return NULL;
@@ -1496,7 +1496,7 @@ _ai_request_thread(gpointer data)
         log_debug("[AI-THREAD] HTTP error response body (%zu bytes): %s",
                   response.size, response.data ? response.data : "NULL");
         /* Try to extract the actual error message from the JSON response */
-        auto_gchar gchar* parsed_error = _parse_error_response(response.data);
+        auto_gchar gchar* parsed_error = ai_parse_error_message(response.data);
         auto_gchar gchar* error_msg = parsed_error ? g_strdup_printf("HTTP %ld: %s", http_code, parsed_error) : g_strdup_printf("HTTP %ld: %s", http_code, response.data && strlen(response.data) > 0 ? response.data : "Unknown error");
         log_error("AI request failed for %s/%s: %s", local_provider_name, local_model, error_msg);
         _aiwin_display_error(user_data, error_msg);
@@ -1505,7 +1505,7 @@ _ai_request_thread(gpointer data)
         log_debug("[AI-THREAD] Raw API response (%zu bytes): %s", response.size, response.data ? response.data : "NULL");
         auto_gchar gchar* response_data = response.data;
         response.data = NULL;
-        auto_gchar gchar* content = _parse_ai_response(response_data);
+        auto_gchar gchar* content = ai_parse_response(response_data);
         if (content) {
             /* Add assistant response to history (under lock) */
             pthread_mutex_lock(&session->lock);

src/ai/ai_client.h

@@ -181,7 +181,7 @@ gboolean ai_fetch_models(const gchar* provider_name, gpointer user_data);
 gboolean ai_models_are_fresh(const gchar* provider_name);
 
 /* ========================================================================
- * Model Parsing (for testing)
+ * Parsing helpers (exposed for testing)
  * ======================================================================== */
 
 /**
@@ -192,6 +192,24 @@ gboolean ai_models_are_fresh(const gchar* provider_name);
  */
 void ai_parse_models_from_json(AIProvider* provider, const gchar* json);
 
+/**
+ * Extract assistant content from an LLM response body. Tries Perplexity
+ * /v1/responses "text" first, falls back to OpenAI "content". Decodes
+ * the \" and \n escape sequences only.
+ * @param response_json The JSON body from the provider
+ * @return Newly allocated content string, or NULL if not found (caller frees)
+ */
+gchar* ai_parse_response(const gchar* response_json);
+
+/**
+ * Extract human-readable error message from an API error envelope.
+ * Expected format: {"error":{"message":"...","type":"...","code":...}}
+ * Decodes \" \n \\ \t escapes.
+ * @param error_json The JSON body of the error response
+ * @return Newly allocated error message, or NULL if not parseable (caller frees)
+ */
+gchar* ai_parse_error_message(const gchar* error_json);
+
 /* ========================================================================
  * Session Management
  * ======================================================================== */