feat: Ollama/gemma4 compat — /init flow, stream filter, safety fixes

provider/openai:
- Fix doubled tool call args (argsComplete flag): Ollama sends complete
  args in the first streaming chunk then repeats them as delta, causing
  doubled JSON and 400 errors in elfs
- Handle fs: prefix (gemma4 uses fs:grep instead of fs.grep)
- Add Reasoning field support for Ollama thinking output

cmd/gnoma:
- Early TTY detection so logger is created with correct destination
  before any component gets a reference to it (fixes slog WARN bleed
  into TUI textarea)

permission:
- Exempt spawn_elfs and agent tools from safety scanner: elf prompt
  text may legitimately mention .env/.ssh/credentials patterns and
  should not be blocked

tui/app:
- /init retry chain: no-tool-calls → spawn_elfs nudge → write nudge
  (ask for plain text output) → TUI fallback write from streamBuf
- looksLikeAgentsMD + extractMarkdownDoc: validate and clean fallback
  content before writing (reject refusals, strip narrative preambles)
- Collapse thinking output to 3 lines; ctrl+o to expand (live stream
  and committed messages)
- Stream-level filter for model pseudo-tool-call blocks: suppresses
  <<tool_code>>...</tool_code>> and <<function_call>>...<tool_call|>
  from entering streamBuf across chunk boundaries
- sanitizeAssistantText regex covers both block formats
- Reset streamFilterClose at every turn start

internal/engine/engine.go

@@ -45,6 +45,11 @@ type Turn struct {
 	Rounds   int               // number of API round-trips
 }
 
+// TurnOptions carries per-turn overrides that apply for a single Submit call.
+type TurnOptions struct {
+	ToolChoice provider.ToolChoiceMode // "" = use provider default
+}
+
 // Engine orchestrates the conversation.
 type Engine struct {
 	cfg     Config
@@ -59,6 +64,9 @@ type Engine struct {
 	// Deferred tool loading: tools with ShouldDefer() are excluded until
 	// the model requests them. Activated on first use.
 	activatedTools map[string]bool
+
+	// Per-turn options, set for the duration of SubmitWithOptions.
+	turnOpts TurnOptions
 }
 
 // New creates an engine.
@@ -124,6 +132,9 @@ func (e *Engine) ContextWindow() *gnomactx.Window {
 // the model should see as context in subsequent turns.
 func (e *Engine) InjectMessage(msg message.Message) {
 	e.history = append(e.history, msg)
+	if e.cfg.Context != nil {
+		e.cfg.Context.AppendMessage(msg)
+	}
 }
 
 // Usage returns cumulative token usage.
@@ -145,4 +156,8 @@ func (e *Engine) SetModel(model string) {
 func (e *Engine) Reset() {
 	e.history = nil
 	e.usage = message.Usage{}
+	if e.cfg.Context != nil {
+		e.cfg.Context.Reset()
+	}
+	e.activatedTools = make(map[string]bool)
 }