chore(release): 1.19.3 [skip ci]

## [1.19.3](https://github.com/asepharyana/zesdex/compare/v1.19.2...v1.19.3) (2026-08-27) ### Bug Fixes * **api:** model Opus pakai URL + API custom dari ~/.claude/settings.json ([1f91b44](https://github.com/asepharyana/zesdex/commit/1f91b447080e7106201d77585edb7d38b74e34bb))
fix(api): model Opus pakai URL + API custom dari ~/.claude/settings.json
2026-08-27 16:49:19 +00:00 · 2026-08-27 23:45:30 +07:00 · 2026-08-27 16:30:31 +00:00 · 2026-08-27 23:26:37 +07:00 · 2026-08-27 23:25:44 +07:00 · 2026-08-27 16:10:48 +00:00
24 changed files with 830 additions and 813 deletions
@@ -7,3 +7,4 @@ package-lock.json
 .superpowers/
 docs/lesson/
 .kilo/
 .hermes/
@@ -1,3 +1,24 @@
 ## [1.19.3](https://github.com/asepharyana/zesdex/compare/v1.19.2...v1.19.3) (2026-08-27)
 ### Bug Fixes
 * **api:** model Opus pakai URL + API custom dari ~/.claude/settings.json ([1f91b44](https://github.com/asepharyana/zesdex/commit/1f91b447080e7106201d77585edb7d38b74e34bb))
 ## [1.19.2](https://github.com/asepharyana/zesdex/compare/v1.19.1...v1.19.2) (2026-08-27)
 ### Performance Improvements
 * **agent:** stabilkan async & parallel — satu runtime, bounded concurrency, isolasi error ([6a98d52](https://github.com/asepharyana/zesdex/commit/6a98d52d54a69f78710d852a69dda3ac0a4ead31))
 ## [1.19.1](https://github.com/asepharyana/zesdex/compare/v1.19.0...v1.19.1) (2026-08-27)
 ### Performance Improvements
 * **agent:** rombak alur AI agent — adaptif, hemat token, self-healing ([eac0443](https://github.com/asepharyana/zesdex/commit/eac0443c4c3b8bfcbefd4bad9554168fb6525b94))
 # [1.19.0](https://github.com/asepharyana/zesdex/compare/v1.18.4...v1.19.0) (2026-08-27)
@@ -4862,7 +4862,7 @@ dependencies = [
 [[package]]
 name = "zesdex-api"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "argon2",
@@ -4885,7 +4885,7 @@ dependencies = [
 [[package]]
 name = "zesdex-application"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "base64",
@@ -4902,7 +4902,7 @@ dependencies = [
 [[package]]
 name = "zesdex-bootstrap"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "chrono",
@@ -4919,7 +4919,7 @@ dependencies = [
 [[package]]
 name = "zesdex-daemon"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "base64",
@@ -4943,7 +4943,7 @@ dependencies = [
 [[package]]
 name = "zesdex-domain"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "base64",
@@ -4959,7 +4959,7 @@ dependencies = [
 [[package]]
 name = "zesdex-gateway"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "axum",
@@ -4986,7 +4986,7 @@ dependencies = [
 [[package]]
 name = "zesdex-grpc"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "axum",
@@ -5003,7 +5003,7 @@ dependencies = [
 [[package]]
 name = "zesdex-infrastructure"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "argon2",
@@ -5051,7 +5051,7 @@ dependencies = [
 [[package]]
 name = "zesdex-tui"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "base64",
@@ -5077,7 +5077,7 @@ dependencies = [
 [[package]]
 name = "zesdex-web"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "axum",
@@ -5097,7 +5097,7 @@ dependencies = [
 [[package]]
 name = "zesdex-ws"
-version = "1.18.3"
+version = "1.19.1"
 dependencies = [
 "anyhow",
 "axum",
@@ -15,7 +15,7 @@ members = [
 ]
 [workspace.package]
-version = "1.19.0"
+version = "1.19.3"
 edition = "2021"
 authors = ["asepharyana <superaseph@gmail.com>"]
@@ -1,250 +0,0 @@
 # Zesdex — Autonomous AI Coding Agent
 Zesdex is an autonomous AI coding agent with a Terminal UI (TUI). It acts as an
 OpenAI/Anthropic-compatible LLM client wrapped in a tool-use harness with **37
 built-in tools** — file operations, git, shell execution, LSP integration, MCP,
 subagent orchestration, and more.
 ```
 ┌──────────────────────────────────────────────────────────────┐
 │                       Mode Selector                            │
 │  TUI (default) ─── Daemon ─── Attach ─── API ─── WS/gRPC/Web │
 └──────────────────────────────────────────────────────────────┘
 ```
 ---
 ## Quick Start
 ```bash
 # Run the TUI (default mode)
 cargo run
 # Run the REST API server
 cargo run -- --api --api-port 8080
 # Run in daemon mode (background + IPC)
 cargo run -- --daemon
 # Attach TUI to a running daemon session
 cargo run -- --attach <session-id>
 # Seed initial data (first run)
 cargo run --bin bootstrap
 ```
 ### Prerequisites
 - **Rust** 1.81+ (edition 2021)
 - **Linux** or **macOS** (Unix domain sockets required for daemon mode)
 - An **API key** for an OpenAI/Anthropic-compatible LLM provider (set via
  settings or environment variable)
 ---
 ## Modes
 | Flag | Mode | Description |
 |------|------|-------------|
 | *(none)* | **TUI** | Full terminal UI with chat, overlays, and agent loop in one process |
 | `--daemon` | **Daemon** | Background daemon with IPC socket; clients attach separately |
 | `--attach <id>` | **Attach** | Connect TUI to an existing daemon session via Unix socket |
 | `--api` | **REST API** | HTTP server with session management and chat endpoints |
 | `--ws` | **WebSocket** | WebSocket server for real-time communication |
 | `--grpc` | **gRPC** | gRPC server for programmatic access |
 | `--web` | **Web** | Serves the web frontend |
 | `--api-port`, `--ws-port`, `--grpc-port`, `--web-port` | *(ports)* | Configure server ports (defaults: 8080, 8081, 50051, 3000) |
 ---
 ## Architecture
 ### Clean Architecture Layering
 ```
 apps/
 ├── domain/          # Pure entities, value objects, repository/service traits
 │                    # Zero framework deps — only serde + chrono + uuid
 ├── application/     # Use-case services (auth, sessions, conversations, memory)
 │                    # Depends only on domain-layer trait interfaces
 ├── infrastructure/  # All I/O: LLM client, IPC, persistence, LSP, MCP, tools
 │                    # Implements domain/application port interfaces
 └── interfaces/      # Entry points
    ├── tui/         # Ratatui terminal UI
    ├── api/         # Axum REST API
    ├── daemon/      # Unix socket daemon + client
    ├── ws/          # WebSocket server
    ├── grpc/        # gRPC server
    └── web/         # Web frontend (static file server)
 ```
 ### Tool System
 37 tools across 9 categories:
 | Category | Tools |
 |----------|-------|
 | **File System** | `read`, `write`, `edit`, `delete`, `dir_list`, `dir_cache_update` |
 | **Shell** | `bash`, `bash_interactive`, `bash_kill`, `bash_output` |
 | **Git** | `git_operator`, `git_cred`, `git_worktree` |
 | **Search** | `search`, `grep`, `glob`, `semantic_search` |
 | **LSP** | `lsp_connect`, `lsp_hover`, `lsp_completion`, `lsp_definition`, `lsp_references`, `lsp_diagnostics`, `lsp_disconnect` |
 | **Memory** | `remember`, `recall`, `forget` |
 | **Workflow** | `spawn_agents`, `spawn_pipeline`, `plan`, `sequential_think`, `hive_mind` |
 | **Utility** | `todo_write`, `todo_finish`, `pong`, `cd` |
 | **Background** | Background bash jobs with `cancel/status/list` |
 Each tool implements the `Tool` trait:
 ```rust
 pub trait Tool: Send + Sync {
    fn name(&self) -> &'static str;
    fn description(&self) -> &'static str;
    fn parameters(&self) -> Value;
    fn run(&self, ctx: &ToolCtx, args: &Value) -> Result<String>;
 }
 ```
 ### Hive Mind Orchestration
 The multi-agent orchestration system compiles a **cognitive cycle plan** per
 task — ordered cycles of parallel processing nodes. Each node has a directive
 and an **access tier** (`read` / `write` / `full`). Node outputs merge into a
 shared collective state in real time, and a final **consensus synthesis**
 produces the unified result.
 - **Auto-trigger**: Complex requests automatically use the hive mind
 - **Manual entry**: The `hive_mind` tool lets the LLM specify cycles explicitly
 - **Live progress**: TUI panel shows each node's status and current tool
 - **Guaranteed docs**: Every convergence writes to `docs/runs/`
 ### IPC Protocol (Daemon Mode)
 ```
 ┌──────────┐   Unix socket    ┌──────────┐
 │  Client   │ ◄──────────────► │  Daemon  │
 │  (TUI)    │   length-prefixed│          │
 └──────────┘   serde_json      └──────────┘
 Frame format: [4-byte BE length][JSON payload]
 ```
 The daemon holds `AppStateRest` and drives the agent loop. Clients are stateless
 renderers that receive full state snapshots after each action.
 ---
 ## Built-in Features
 | Feature | Description |
 |---------|-------------|
 | **LLM Provider** | OpenAI/Anthropic-compatible API (streaming + non-streaming) with automatic retry and fallback |
 | **Tool Harness** | Safety-gated tool execution with graduated review checks |
 | **Subagents** | Auto-inline review, background test-gen, arch-review, security-review |
 | **OAuth 2.0** | PKCE flow for LLM provider authentication |
 | **MCP** | Model Context Protocol server management (stdio + HTTP transport) |
 | **LSP** | Language Server Protocol integration (completion, hover, diagnostics, references) |
 | **Session Mgmt** | SQLite-persisted sessions with lock-based concurrency control |
 | **Memory** | File-based memory system with frontmatter metadata |
 | **Edit Log** | Append-only edit history with configurable retention |
 | **Rate Limiting** | Sliding-window per-client rate limiter |
 | **JWT Auth** | HS256 JWT access/refresh tokens (API mode) |
 | **Password Auth** | Argon2 password hashing with pepper |
 | **OAuth Loopback** | Localhost HTTP server for OAuth redirect capture |
 | **Background Jobs** | Long-running shell jobs with cancellation and output collection |
 | **Settings** | JSON-persisted settings with hot-reload |
 ---
 ## TUI Overlays
 16 overlays accessible from the terminal UI:
 | Overlay | Purpose |
 |---------|---------|
 | Chat Input | Main input bar with autocomplete |
 | Bash Panel | Interactive shell panel |
 | File Editor | Built-in file editor |
 | Effort Selector | LLM reasoning effort selector |
 | Help | Keybindings reference |
 | Key Input | Custom key binding configuration |
 | Learning | Lesson viewer |
 | Loading | Generating spinner |
 | MCP Manager | MCP server management |
 | Model Selector | LLM model picker |
 | Quit Confirm | Exit confirmation dialog |
 | Rewind | Message/history rewind |
 | Settings | Settings panel |
 | Todo | Task/TODO list |
 | Usage | Token usage statistics |
 | Workflow | Hive-mind node progress |
 ---
 ## Data & Persistence
 All data lives under the platform's data directory (`~/.local/share/zesdex/`):
 ```
 ~/.local/share/zesdex/
 ├── settings.json           # User settings (provider, model, keys)
 ├── app_config.json         # Provider definitions (endpoints, env vars)
 ├── sessions/               # Chat sessions (one subdirectory per session)
 │   └── <uuid>/
 │       ├── session.json    # Session metadata
 │       ├── messages.jsonl  # Message log
 │       └── .lock           # Session lock file
 └── memories/               # Memory files with frontmatter metadata
    └── *.md
 ```
 ---
 ## Development
 ```bash
 # Build all crates
 cargo build
 # Run all unit tests (8 tests across 11 crates)
 cargo test
 # Run clippy linting
 cargo clippy --all-targets
 # Run with verbose logging
 RUST_LOG=debug cargo run
 ```
 ### Workspace Crates
 | Crate | Path | Layer |
 |-------|------|-------|
 | `zesdex-domain` | `apps/domain/` | Pure domain entities & traits |
 | `zesdex-application` | `apps/application/` | Use-case services |
 | `zesdex-infrastructure` | `apps/infrastructure/` | All I/O & tool implementations |
 | `zesdex-tui` | `apps/interfaces/tui/` | Ratatui terminal interface |
 | `zesdex-api` | `apps/interfaces/api/` | Axum REST API |
 | `zesdex-daemon` | `apps/interfaces/daemon/` | Unix socket daemon |
 | `zesdex-ws` | `apps/interfaces/ws/` | WebSocket server |
 | `zesdex-grpc` | `apps/interfaces/grpc/` | gRPC server |
 | `zesdex-web` | `apps/interfaces/web/` | Web frontend |
 | `zesdex-gateway` | `apps/gateway/` | CLI entry point & dispatcher |
 | `zesdex-bootstrap` | `apps/bootstrap/` | Initial data seeder |
 ### Code Map
 Detailed architecture documentation is in `docs/CODEMAPS/`:
 | File | Covers |
 |------|--------|
 | `docs/CODEMAPS/architecture.md` | System layout, process modes, data flow |
 | `docs/CODEMAPS/backend.md` | Provider, OAuth, IPC, workflow engine, MCP, LSP, review |
 | `docs/CODEMAPS/frontend.md` | TUI render pipeline, 16 overlays, toasts, input handling |
 | `docs/CODEMAPS/data.md` | Persistence, SQLite msglog, memory files, settings/config |
 | `docs/CODEMAPS/dependencies.md` | All Rust crates and external services |
 ---
 ## License
 See `CHANGELOG.md` for release history.
@@ -1,57 +0,0 @@
 //! Mandatory explore phase — spawns parallel subagents to discover context
 //! before the main agent begins its turn.
 //!
 //! # Flow
 //!
 //! Before the main agent's LLM loop, [`ExploreService::explore`] dispatches
 //! at least 3 subagents in parallel (code-structure scan, symbol-index query,
 //! semantic-context search). Their findings are consolidated into a single
 //! system message that is prepended to the conversation.
 //!
 //! # Why mandatory
 //!
 //! Without structured exploration the main agent works from an empty context
 //! window. The explore phase guarantees that every turn starts with a compact
 //! snapshot of what the codebase contains and where relevant code lives.
 use anyhow::Result;
 use std::collections::VecDeque;
 use std::future::Future;
 use std::pin::Pin;
 use std::sync::{Arc, Mutex};
 use zesdex_domain::agent::TurnEvent;
 /// The consolidated output of an explore phase — a set of system-level
 /// context messages injected before the main agent prompt.
 #[derive(Debug, Clone)]
 pub struct ExploreOutput {
    /// One or more system messages summarising what the explore subagents
    /// discovered. Prepended to the conversation by the turn service.
    pub context_messages: Vec<String>,
    /// Short human-readable summary of what was explored.
    pub summary: String,
 }
 /// Service trait for the mandatory pre-turn exploration phase.
 ///
 /// Implementors spawn ≥3 parallel subagents, each analysing a different
 /// aspect of the workspace, and return a consolidated summary.
 ///
 /// # Object safety
 ///
 /// This trait is `dyn`-safe — it returns `Pin<Box<dyn Future>>` so it can
 /// be stored as `Arc<dyn ExploreService>`.
 pub trait ExploreService: Send + Sync {
    /// Run the explore phase.
    ///
    /// `query` — the user's current input phrase.
    /// `workspace_root` — absolute path to the workspace root.
    /// `turn_events` — shared event queue for TUI updates.
    /// Returns structured context messages and a summary blob.
    fn explore<'a>(
        &'a self,
        query: &'a str,
        workspace_root: &'a str,
        turn_events: &'a Arc<Mutex<VecDeque<TurnEvent>>>,
    ) -> Pin<Box<dyn Future<Output = Result<ExploreOutput>> + Send + 'a>>;
 }
@@ -19,8 +19,6 @@ pub trait AgentTurnService: Send + Sync {
    fn run_turn(&self, params: AgentTurnParams) -> impl Future<Output = Result<()>> + Send;
 }
 pub mod explore;
 pub mod turn_service;
 pub use explore::{ExploreOutput, ExploreService};
 pub use turn_service::{compact_messages_with_ai, AgentTurnServiceImpl};
@@ -7,12 +7,29 @@ use zesdex_domain::agent::{AgentTurnParams, TurnEvent};
 use zesdex_domain::core::{ChatMessage, StreamEvent, ToolDef};
 use zesdex_domain::main_agent_prompt;
-use super::{ExploreService, ToolExecutor};
+use super::ToolExecutor;
 use crate::ports::ProviderService;
 /// Maximum tool-call iterations per agent turn before forcing termination.
 const MAX_TURN_ITERATIONS: u32 = 50;
 /// Maximum number of consecutive identical tool errors before the loop
 /// injects a recovery note and forces a different approach.
 const MAX_CONSECUTIVE_TOOL_ERRORS: usize = 3;
 /// Total tool-call errors tolerated per turn before the loop is stopped.
 const MAX_TOTAL_TOOL_ERRORS: usize = 8;
 /// Ceiling for a single tool-result message inserted into context.
 ///
 /// Tool outputs can be huge (read / semantic_search). Truncating keeps the
 /// context window from exploding while preserving the important head.
 const TOOL_OUTPUT_MAX_CHARS: usize = 12_000;
 /// Total conversation characters that trigger auto-compaction before the
 /// next LLM call.
 const AUTO_COMPACT_CHARS: usize = 60_000;
 // ---------------------------------------------------------------------------
 // Helper: push a TurnEvent onto the shared queue.
 // ---------------------------------------------------------------------------
@@ -51,6 +68,89 @@ fn make_stream_callback(
    })
 }
 // ---------------------------------------------------------------------------
 // Helper: truncate a long tool output before it enters the conversation
 // context. Preserves the head and appends a clear truncation marker.
 // ---------------------------------------------------------------------------
 fn truncate_tool_output(output: String) -> String {
    if output.len() <= TOOL_OUTPUT_MAX_CHARS {
        return output;
    }
    let mut result: String = output.chars().take(TOOL_OUTPUT_MAX_CHARS).collect();
    result.push_str(&format!(
        "\n...[truncated {} chars]",
        output.len() - TOOL_OUTPUT_MAX_CHARS
    ));
    result
 }
 // ---------------------------------------------------------------------------
 // Helper: adaptive generation parameters.
 // ---------------------------------------------------------------------------
 /// Pick a `max_tokens` budget for the turn's next LLM call based on the
 /// length of the user's request. Short requests need far fewer tokens than
 /// the current hardcoded 4096 — big savings on small tasks.
 fn adaptive_max_tokens(request_len: usize) -> u32 {
    if request_len <= 80 {
        800
    } else if request_len <= 400 {
        1600
    } else {
        4096
    }
 }
 /// Sum the character length of the conversation (user + assistant +
 /// tool content) as a cheap proxy for context size.
 fn conversation_chars(messages: &[ChatMessage]) -> usize {
    messages
        .iter()
        .map(|m| m.content.as_deref().map(str::len).unwrap_or(0))
        .sum()
 }
 /// Track repeated tool-call errors so the loop can recover instead of
 /// burning iterations retrying the same failing tool.
 #[derive(Default)]
 struct ErrorTracker {
    consecutive: usize,
    total: usize,
    last_tool: String,
    last_error: String,
 }
 impl ErrorTracker {
    fn record(&mut self, tool_name: &str, error: &str, messages: &mut Vec<ChatMessage>) {
        if self.last_tool == tool_name {
            self.consecutive += 1;
        } else {
            self.consecutive = 1;
        }
        self.last_tool = tool_name.to_string();
        self.last_error = error.to_string();
        self.total += 1;
        // Inject a recovery note once the same tool keeps failing.
        if self.consecutive >= MAX_CONSECUTIVE_TOOL_ERRORS
            && !messages.iter().any(|m| {
                m.content
                    .as_deref()
                    .is_some_and(|c| c.contains("[System note]"))
            })
        {
            messages.push(ChatMessage::system(
                zesdex_domain::agent::prompt::error_recovery_note(tool_name, error),
            ));
        }
    }
    fn should_stop(&self) -> bool {
        self.consecutive >= MAX_CONSECUTIVE_TOOL_ERRORS * 2 || self.total >= MAX_TOTAL_TOOL_ERRORS
    }
 }
 // ---------------------------------------------------------------------------
 // Helper: execute a single tool call, push events, return the result string.
 // ---------------------------------------------------------------------------
@@ -71,6 +171,7 @@ async fn execute_tool_call<T: ToolExecutor>(
    };
    let is_error = output.starts_with("Error:");
    let output = truncate_tool_output(output);
    push_event(
        turn_events,
@@ -108,19 +209,19 @@ fn emit_usage(turn_events: &Arc<Mutex<VecDeque<TurnEvent>>>, usage: Option<(u64,
 /// Service implementation for executing an agent turn asynchronously.
 ///
-/// # Explore phase
+/// The turn loop is adaptive and token-aware:
-///
+/// - No mandatory explore phase — the *agent* decides when to call the
-/// Before the main LLM loop begins, [`AgentTurnServiceImpl`] runs a mandatory
+///   `explore_codebase` tool (see the main prompt), so simple queries skip
-/// explore phase that spawns ≥3 parallel subagents (code structure, symbol
+///   exploration entirely.
-/// index, semantic context) and injects their consolidated findings as a
+/// - `max_tokens` / `temperature` adapt to the request length and phase.
-/// system message. See [`ExploreService`] for the trait contract.
+/// - Repeated tool errors trigger a system recovery note and eventually
 ///   stop the loop instead of burning iterations.
 /// - Tool outputs are truncated before entering context.
 /// - Oversized histories are auto-compacted before the next LLM call.
 pub struct AgentTurnServiceImpl<P: ProviderService, T: ToolExecutor> {
    provider: Arc<P>,
    tool_executor: Arc<T>,
    tool_defs: Vec<ToolDef>,
    /// Optional explore-phase service. When `Some`, the explore phase runs
    /// before every turn; when `None` it is skipped (tests, daemon mode).
    explore_service: Option<Arc<dyn ExploreService>>,
 }
 impl<P: ProviderService, T: ToolExecutor> AgentTurnServiceImpl<P, T> {
@@ -129,19 +230,9 @@ impl<P: ProviderService, T: ToolExecutor> AgentTurnServiceImpl<P, T> {
            provider,
            tool_executor,
            tool_defs,
            explore_service: None,
        }
    }
    /// Attach an optional explore-phase service.
    ///
    /// When set, every call to `run_turn` will first run the explore phase
    /// and inject the consolidated context as a system message.
    pub fn with_explore(mut self, service: Arc<dyn ExploreService>) -> Self {
        self.explore_service = Some(service);
        self
    }
    /// Execute a single LLM call with the current message list, handling
    /// streaming events and error reporting.
    async fn call_llm(
@@ -149,6 +240,8 @@ impl<P: ProviderService, T: ToolExecutor> AgentTurnServiceImpl<P, T> {
        messages: &[ChatMessage],
        abort: &Arc<AtomicBool>,
        turn_events: &Arc<Mutex<VecDeque<TurnEvent>>>,
        max_tokens: u32,
        temperature: f32,
    ) -> Result<(ChatMessage, Option<(u64, u64)>), String> {
        let on_event = make_stream_callback(abort, turn_events);
@@ -156,13 +249,39 @@ impl<P: ProviderService, T: ToolExecutor> AgentTurnServiceImpl<P, T> {
            .chat_stream(
                messages,
                Some(self.tool_defs.clone()),
-                Some(4096),
+                Some(max_tokens),
-                Some(0.7),
+                Some(temperature),
                on_event,
            )
            .await
            .map_err(|e| format!("LLM error: {e}"))
    }
    /// Auto-compact the history in place if it exceeds the threshold.
    ///
    /// Runs at most once per turn. Skips the synthetic system prompt that
    /// this service inserts at index 0.
    async fn auto_compact_if_needed(&self, messages: &mut Vec<ChatMessage>) {
        if conversation_chars(messages) <= AUTO_COMPACT_CHARS {
            return;
        }
        // Keep the system prompt (index 0) out of compaction.
        let sys = messages[0].clone();
        let mut rest: Vec<ChatMessage> = messages.drain(1..).collect();
        let before = rest.len();
        if let Err(e) = super::compact_messages_with_ai(&mut rest, self.provider.as_ref()).await {
            warn!("auto-compact failed (non-fatal): {e}");
        }
        info!(
            "auto-compacted history: {} messages -> {}",
            before,
            rest.len()
        );
        let mut rebuilt = Vec::with_capacity(rest.len() + 1);
        rebuilt.push(sys);
        rebuilt.extend(rest);
        *messages = rebuilt;
    }
 }
 impl<P: ProviderService, T: ToolExecutor> super::AgentTurnService for AgentTurnServiceImpl<P, T> {
@@ -173,73 +292,26 @@ impl<P: ProviderService, T: ToolExecutor> super::AgentTurnService for AgentTurnS
            params.model
        );
        // ── Phase 0: Mandatory explore ──────────────────────────────────
        // Spawn ≥3 parallel subagents to discover code structure, symbols,
        // and semantic context. The consolidated summary is injected as a
        // system message before the main agent prompt.
        if let Some(ref explorer) = self.explore_service {
            // Determine workspace root from the first message's context or
            // the first workspace root in params.
            let user_query = params
                .messages
                .last()
                .map(|m| m.content.clone().unwrap_or_default())
                .unwrap_or_default();
            let workspace_root = params
                .workspace_roots
                .first()
                .map(|p| p.to_string_lossy().to_string())
                .unwrap_or_else(|| ".".to_string());
            push_event(
                &params.turn_events,
                TurnEvent::SystemNote {
                    kind: "info".into(),
                    message: "🔍 Exploring codebase structure...".into(),
                },
            );
            match explorer
                .explore(&user_query, &workspace_root, &params.turn_events)
                .await
            {
                Ok(output) => {
                    // Insert each context message as a system message.
                    // They go at index 0 and are removed after the turn
                    // like the main agent prompt.
                    for ctx_msg in &output.context_messages {
                        params
                            .messages
                            .insert(0, ChatMessage::system(ctx_msg.clone()));
                    }
                    info!(
                        "Explore phase complete: {} context messages, {}",
                        output.context_messages.len(),
                        output.summary
                    );
                }
                Err(e) => {
                    warn!("Explore phase failed (non-fatal): {e}");
                    push_event(
                        &params.turn_events,
                        TurnEvent::SystemNote {
                            kind: "warn".into(),
                            message: format!("Explore phase failed: {e}"),
                        },
                    );
                }
            }
        }
        // Insert system prompt at position 0 once and keep it there for the
        // entire turn, avoiding per-iteration clones of the full message list.
        // It is removed before emitting the Compacted event so persistence
        // does not store the prompt redundantly.
        params
            .messages
            .insert(0, ChatMessage::system(main_agent_prompt()));
        let original_count = params.messages.len();
        // Estimate request complexity from the last user message.
        let request_len = params
            .messages
            .last()
            .and_then(|m| m.content.as_deref())
            .map(str::len)
            .unwrap_or(0);
        let mut errors = ErrorTracker::default();
        // Track whether the previous call produced tool calls — used to
        // lower temperature once the agent starts producing a final answer.
        let mut saw_tool_calls = false;
        for iteration in 0..MAX_TURN_ITERATIONS {
            // ── Check abort flag ────────────────────────────────────────
            if params.abort.load(Ordering::SeqCst) {
@@ -254,15 +326,40 @@ impl<P: ProviderService, T: ToolExecutor> super::AgentTurnService for AgentTurnS
                break;
            }
            if errors.should_stop() {
                push_event(
                    &params.turn_events,
                    TurnEvent::SystemNote {
                        kind: "warn".into(),
                        message: "Stopping: repeated tool errors without progress".into(),
                    },
                );
                break;
            }
            debug!("agent turn iteration {iteration}");
            // ── Auto-compact oversized history before the LLM call ─────
            self.auto_compact_if_needed(&mut params.messages).await;
            // ── Adaptive generation parameters ─────────────────────────
            let max_tokens = adaptive_max_tokens(request_len);
            // Lower temperature while the agent is still choosing tools to
            // keep tool selection deterministic; raise it for the final
            // free-form answer.
            let temperature = if saw_tool_calls { 0.2 } else { 0.7 };
            // ── Stream start + call LLM ─────────────────────────────────
            push_event(&params.turn_events, TurnEvent::StreamStart);
            // Uses params.messages directly (sys_msg[0] already in place
            // from the insert above) — no per-iteration clone needed.
            let result = self
-                .call_llm(&params.messages, &params.abort, &params.turn_events)
+                .call_llm(
                    &params.messages,
                    &params.abort,
                    &params.turn_events,
                    max_tokens,
                    temperature,
                )
                .await;
            match result {
@@ -283,6 +380,7 @@ impl<P: ProviderService, T: ToolExecutor> super::AgentTurnService for AgentTurnS
                        break;
                    }
                    saw_tool_calls = true;
                    params.messages.push(assistant_msg);
                    // ── Execute each tool call ──────────────────────────
@@ -290,6 +388,9 @@ impl<P: ProviderService, T: ToolExecutor> super::AgentTurnService for AgentTurnS
                        let output =
                            execute_tool_call(self.tool_executor.as_ref(), &params.turn_events, tc)
                                .await;
                        if output.starts_with("Error:") {
                            errors.record(&tc.function.name, &output, &mut params.messages);
                        }
                        params
                            .messages
                            .push(ChatMessage::tool(tc.id.clone(), output));
@@ -370,3 +471,69 @@ pub async fn compact_messages_with_ai<P: ProviderService>(
        }
    }
 }
 // ---------------------------------------------------------------------------
 // Tests
 // ---------------------------------------------------------------------------
 #[cfg(test)]
 mod tests {
    use super::*;
    #[test]
    fn truncate_short_output_is_unchanged() {
        let out = "short".to_string();
        assert_eq!(truncate_tool_output(out.clone()), out);
    }
    #[test]
    fn truncate_long_output_preserves_head_and_marks_cut() {
        let long = "x".repeat(TOOL_OUTPUT_MAX_CHARS + 500);
        let truncated = truncate_tool_output(long.clone());
        assert!(truncated.len() < long.len());
        assert!(truncated.contains("...[truncated"));
        assert!(truncated.starts_with("xxx"));
    }
    #[test]
    fn adaptive_max_tokens_scales_with_request_len() {
        assert_eq!(adaptive_max_tokens(10), 800);
        assert_eq!(adaptive_max_tokens(200), 1600);
        assert_eq!(adaptive_max_tokens(5000), 4096);
    }
    #[test]
    fn error_tracker_injects_recovery_note_after_repeats() {
        let mut tracker = ErrorTracker::default();
        let mut messages: Vec<ChatMessage> = Vec::new();
        tracker.record("read", "Error: File not found", &mut messages);
        tracker.record("read", "Error: File not found", &mut messages);
        assert!(!tracker.should_stop());
        // Third consecutive failure → recovery note injected.
        tracker.record("read", "Error: File not found", &mut messages);
        assert!(messages.iter().any(|m| m
            .content
            .as_deref()
            .is_some_and(|c| c.contains("[System note]"))));
    }
    #[test]
    fn error_tracker_stops_after_too_many_errors() {
        let mut tracker = ErrorTracker::default();
        let mut messages: Vec<ChatMessage> = Vec::new();
        for i in 0..MAX_TOTAL_TOOL_ERRORS {
            tracker.record("bash", &format!("Error: boom {i}"), &mut messages);
        }
        assert!(tracker.should_stop());
    }
    #[test]
    fn conversation_chars_sums_content_only() {
        let messages = vec![
            ChatMessage::system("sys".to_string()),
            ChatMessage::user("hello world".to_string()),
            ChatMessage::tool("id".to_string(), "output".to_string()),
        ];
        assert_eq!(conversation_chars(&messages), 3 + 11 + 6);
    }
 }
@@ -52,5 +52,5 @@ pub use cms::{
 pub use agent::{
    turn_service::{compact_messages_with_ai, AgentTurnServiceImpl},
-    AgentTurnService, ExploreOutput, ExploreService, ToolExecutor,
+    AgentTurnService, ToolExecutor,
 };
@@ -22,12 +22,20 @@ pub fn main_agent_prompt() -> String {
 You are Zesdex, an AI coding assistant. You have access to various tools \
 via native function calling to help the user.
 TOKEN BUDGET — BE EFFICIENT:
 - For simple/factual questions, answer directly. Do NOT call tools.
 - For complex or unfamiliar code tasks, call `explore_codebase` ONCE at the \
 start to locate relevant code, then work from that context.
 - Keep tool usage minimal: prefer `grep`/`glob`/`read` for targeted lookups; \
 avoid re-reading files you already have in context.
 - Keep responses concise; do not repeat tool output verbatim.
 CRITICAL DIRECTIVES & PRIORITY HIERARCHY:
 1. WORKFLOW FIRST: For any multi-step, complex, or non-trivial task, \
 you MUST prioritise using `workflow_run` (to construct and execute a \
 multi-phase YAML workflow) or `hive_mind` (to orchestrate parallel \
 autonomous agents). Workflows are your primary strategy.
-2. PLANNING & TODOS: Use `plan_enter` to establish high-level \
+2. PLANNING & TODOs: Use `plan_enter` to establish high-level \
 architectural plans and `todowrite` to maintain granular task checklists.
 3. REASONING: Use `seq_think` for deep step-by-step analysis.
 4. TOOL EXECUTION: Execute individual tools (file edits, terminal commands) \
@@ -70,6 +78,35 @@ executed, and modified files. Format as a clear bulleted list."
        .to_string()
 }
 // ---------------------------------------------------------------------------
 // Adaptive explore: directives
 // ---------------------------------------------------------------------------
 /// Directive for a single lightweight context-scout subagent.
 pub fn explore_scout_directive() -> String {
    "\
 You are a codebase context scout. \
 Given the workspace root, quickly locate the code that is most relevant \
 to the user's request: \
 1. Run semantic_search once with the user's key terms. \
 2. Read up to the 3 most relevant files (use grep for symbols if needed). \
 3. Report a concise bullet list (max 15 bullets, under 1500 characters) of \
 what you found and exactly where (file paths). \
 Do NOT rebuild the index. Do NOT enumerate unrelated files. Be brief."
        .to_string()
 }
 /// Build a system note injected after repeated tool errors to steer the
 /// agent toward an alternative approach instead of retrying the same call.
 pub fn error_recovery_note(tool_name: &str, last_error: &str) -> String {
    format!(
        "\
 [System note] The tool `{tool_name}` failed repeatedly with: \"{last_error}\". \
 Try an alternative approach (verify paths, correct arguments, use a \
 different tool, or finish without this tool). Do NOT retry the same call."
    )
 }
 #[cfg(test)]
 mod tests {
    use super::*;
@@ -89,4 +126,18 @@ mod tests {
        assert!(prompt.contains("/home"));
        assert!(prompt.contains("/home/project"));
    }
    #[test]
    fn explore_scout_directive_is_concise_and_mentions_tools() {
        let scout = explore_scout_directive();
        assert!(scout.contains("scout"));
        assert!(scout.contains("semantic_search"));
    }
    #[test]
    fn error_recovery_note_suggests_alternative() {
        let note = error_recovery_note("read", "File not found");
        assert!(note.contains("read"));
        assert!(note.contains("alternative"));
    }
 }
@@ -1,304 +1,141 @@
-//! Mandatory explore phase — spawns ≥3 parallel subagents to discover
+//! `explore_codebase` tool — lazy, agent-initiated codebase exploration.
-//! codebase context before every agent turn, visible in the TUI workflow tab.
+//!
 //! The main agent decides (via the system prompt) when it needs codebase
 //! context. Unlike the old mandatory explore phase (which ran 3 subagents on
 //! every turn regardless of the question), this tool is invoked only when the
 //! agent judges it necessary — saving tokens on trivial queries while keeping
 //! context available for complex tasks.
 //!
 //! # Flow
 //!
-//! `ExploreServiceImpl::explore()` →
+//! `ExploreCodebase::run` →
-//!
+//! 1. Parse the user's goal / target from args.
-//! 1. Push `WorkflowAgentUpdate { Pending }` for each agent onto the turn-event
+//! 2. Resolve subagent provider credentials from settings.
-//!    queue so the TUI workflow tab shows all 3.
+//! 3. Spawn a single "context scout" subagent (read-only, semantic_search +
-//! 2. Spawn **Code Structure** subagent (thread + tokio runtime).
+//!    read of up to 3 relevant files).
-//! 3. Spawn **Symbol Index** subagent (thread + tokio runtime).
+//! 4. Join the result and return a concise bullet summary as a tool message.
 //! 4. Spawn **Semantic Context** subagent (thread + tokio runtime).
 //! 5. Join all handles via `spawn_blocking`.
 //! 6. Push `Completed` / `Failed` events for each agent.
 //! 7. Consolidate findings into a system message → return.
 use anyhow::Result;
 use serde_json::{json, Value};
 use tracing::{info, warn};
 use zesdex_domain::agent::prompt::explore_scout_directive;
 use zesdex_domain::cms::{AppConfigRepository, SettingsRepository};
 use zesdex_domain::core::Store;
 use crate::persistence::{JsonAppConfigRepository, JsonSettingsRepository};
 use crate::subagent::context::SubagentContext;
 use crate::subagent::division::AccessTier;
 use crate::subagent::engine::run_agent;
-use crate::tools::ToolCtx;
+use crate::tools::{Tool, ToolCtx};
 use anyhow::{Context, Result};
 use std::collections::VecDeque;
 use std::future::Future;
 use std::pin::Pin;
 use std::sync::{Arc, Mutex};
 use std::thread;
 use tracing::{info, warn};
 use zesdex_application::agent::{ExploreOutput, ExploreService};
 use zesdex_domain::agent::{AgentStatus, TurnEvent};
-/// Number of parallel explore subagents.
+/// Maximum characters of the scout's final output to keep in context.
-const EXPLORE_AGENT_COUNT: usize = 3;
+/// The scout is directed to stay under 1500 chars, but this ceiling protects
 /// against rogue output.
 const EXPLORE_OUTPUT_MAX_CHARS: usize = 4000;
-/// IDs for each explore agent (shown in the workflow tab).
+/// `explore_codebase` tool — ask a read-only context-scout subagent to
-const EXPLORE_IDS: [&str; 3] = ["explore-structure", "explore-symbols", "explore-context"];
+/// locate relevant code for the current task.
 pub struct ExploreCodebase;
-/// Display names for the TUI workflow tab.
+impl Tool for ExploreCodebase {
-const EXPLORE_LABELS: [&str; 3] = [
+    fn name(&self) -> &'static str {
-    "📁 Code Structure",
+        "explore_codebase"
    "🔣 Symbol Index",
    "🔍 Semantic Context",
 ];
 /// Directives for each explore subagent.
 const EXPLORE_DIRECTIVES: [&str; 3] = [
    // Agent 0: Code Structure
    "You are a codebase structure explorer.\n\
     1. List all top-level directories and files in the workspace root.\n\
     2. Read Cargo.toml, package.json, or pyproject.toml at the root.\n\
     3. List the apps/ or src/ directory contents.\n\
     4. Identify main entry points (main.rs, main.py, index.ts, etc.).\n\
     5. Count files by extension type.\n\
     Use the ls_dir, read, grep, and glob tools. Be concise.",
    // Agent 1: Symbol Index
    "You are a symbol index explorer.\n\
     1. Call the 'rebuild_index' tool to rebuild the symbol index.\n\
     2. Call the 'list_symbols' tool with max_results: 100.\n\
     3. Identify public APIs, entry points, and key types.\n\
     4. Group symbols by language and kind.\n\
     Be concise. Report what symbols exist and where they live.",
    // Agent 2: Semantic Context
    "You are a semantic context explorer.\n\
     1. Call the 'rebuild_index' tool to ensure the index is fresh.\n\
     2. Search for symbols related to the user's query using semantic_search.\n\
     3. Search for config files, env variables, and settings.\n\
     4. Search for test files and test patterns.\n\
     Be concise. Report relevant code areas for the task.\n\
     Use the semantic_search, grep, glob, and read tools.",
 ];
 // ---------------------------------------------------------------------------
 //  Credentials
 // ---------------------------------------------------------------------------
 /// LLM credentials for explore subagents.
 pub struct Credentials {
    pub base_url: String,
    pub api_key: String,
    pub model: String,
 }
 // ---------------------------------------------------------------------------
 //  ExploreServiceImpl — implements the application-layer trait
 // ---------------------------------------------------------------------------
 /// Concrete [`ExploreService`] that the turn service calls.
 ///
 /// Owns a shared `ToolCtx` and LLM credentials. Each call to `explore()`
 /// spawns 3 subagents in parallel with TUI workflow-tab visibility.
 pub struct ExploreServiceImpl {
    tool_ctx: ToolCtx,
    credentials: Credentials,
 }
 impl ExploreServiceImpl {
    pub fn new(tool_ctx: ToolCtx, credentials: Credentials) -> Self {
        ExploreServiceImpl {
            tool_ctx,
            credentials,
        }
    }
 }
-impl ExploreService for ExploreServiceImpl {
+    fn description(&self) -> &'static str {
-    fn explore<'a>(
+        "Explore the codebase to locate code relevant to a task. Use this \
-        &'a self,
+         once at the start of complex or unfamiliar tasks (implementing a \
-        query: &'a str,
+         feature, fixing a bug, refactoring, navigating a large repo). \
-        workspace_root: &'a str,
+         Do NOT use for simple factual questions about the current \
-        turn_events: &'a Arc<Mutex<VecDeque<TurnEvent>>>,
+         conversation."
-    ) -> Pin<Box<dyn Future<Output = Result<ExploreOutput>> + Send + 'a>> {
+    }
        Box::pin(async move {
            let context = run_explore_phase(
                query,
                workspace_root,
                &self.tool_ctx,
                &self.credentials,
                turn_events,
            )
            .await?;
-            Ok(ExploreOutput {
+    fn parameters(&self) -> Value {
-                context_messages: vec![context],
+        json!({
-                summary: format!("{EXPLORE_AGENT_COUNT} explore agents dispatched"),
+            "type": "object",
-            })
+            "properties": {
                "goal": {
                    "type": "string",
                    "description": "The task or question to explore for"
                }
            },
            "required": ["goal"]
        })
    }
 }
-// ---------------------------------------------------------------------------
+    fn run(&self, ctx: &ToolCtx, args: &Value) -> Result<String> {
-//  Helpers for pushing workflow events
+        let goal = args
-// ---------------------------------------------------------------------------
+            .get("goal")
            .and_then(|v| v.as_str())
            .unwrap_or("")
            .trim()
            .to_string();
-fn push_event(events: &Arc<Mutex<VecDeque<TurnEvent>>>, event: TurnEvent) {
+        if goal.is_empty() {
-    if let Ok(mut q) = events.lock() {
+            return Err(anyhow::anyhow!("missing non-empty 'goal'"));
        q.push_back(event);
    }
 }
 fn emit_pending(events: &Arc<Mutex<VecDeque<TurnEvent>>>, agent_id: &str, display: &str) {
    push_event(
        events,
        TurnEvent::WorkflowAgentUpdate {
            agent_id: agent_id.to_string(),
            agent_name: display.to_string(),
            status: AgentStatus::Pending,
        },
    );
 }
 fn emit_running(events: &Arc<Mutex<VecDeque<TurnEvent>>>, agent_id: &str, display: &str) {
    push_event(
        events,
        TurnEvent::WorkflowAgentUpdate {
            agent_id: agent_id.to_string(),
            agent_name: display.to_string(),
            status: AgentStatus::Running,
        },
    );
 }
 fn emit_completed(events: &Arc<Mutex<VecDeque<TurnEvent>>>, agent_id: &str, display: &str) {
    push_event(
        events,
        TurnEvent::WorkflowAgentUpdate {
            agent_id: agent_id.to_string(),
            agent_name: display.to_string(),
            status: AgentStatus::Completed,
        },
    );
 }
 fn emit_failed(events: &Arc<Mutex<VecDeque<TurnEvent>>>, agent_id: &str, display: &str, msg: &str) {
    push_event(
        events,
        TurnEvent::WorkflowAgentUpdate {
            agent_id: agent_id.to_string(),
            agent_name: display.to_string(),
            status: AgentStatus::Failed(msg.to_string()),
        },
    );
 }
 // ---------------------------------------------------------------------------
 //  Core orchestration
 // ---------------------------------------------------------------------------
 /// Spawn `EXPLORE_AGENT_COUNT` subagents in parallel, emit workflow events
 /// for the TUI tab, join, and consolidate.
 async fn run_explore_phase(
    query: &str,
    workspace_root: &str,
    tool_ctx: &ToolCtx,
    credentials: &Credentials,
    turn_events: &Arc<Mutex<VecDeque<TurnEvent>>>,
 ) -> Result<String> {
    // ── 1. Emit Pending for all agents (appears instantly in workflow tab) ─
    for (i, &id) in EXPLORE_IDS.iter().enumerate() {
        emit_pending(turn_events, id, EXPLORE_LABELS[i]);
    }
    // ── 2. Prepare directives ───────────────────────────────────────────
    let mut directives: Vec<String> = Vec::with_capacity(EXPLORE_AGENT_COUNT);
    for (i, &d) in EXPLORE_DIRECTIVES.iter().enumerate() {
        let mut d = d.to_string();
        if i == 2 {
            d.push_str(&format!("\n\nThe user's current query is: \"{query}\""));
        }
        d.push_str(&format!("\n\nWorkspace root: {workspace_root}"));
        directives.push(d);
    }
-    // ── 3. Spawn all agents on threads ──────────────────────────────────
+        info!("explore_codebase: {goal}");
    let mut handles: Vec<(usize, thread::JoinHandle<Result<String>>)> =
        Vec::with_capacity(EXPLORE_AGENT_COUNT);
-    for i in 0..EXPLORE_AGENT_COUNT {
+        let store = Store::new();
-        emit_running(turn_events, EXPLORE_IDS[i], EXPLORE_LABELS[i]);
+        let settings = JsonSettingsRepository::new()
            .load(&store.base_dir)
            .unwrap_or_default();
        let app_config = JsonAppConfigRepository::new()
            .load(&store.base_dir)
            .unwrap_or_default();
-        let ctx = SubagentContext::new(
+        let (provider, model) =
-            directives[i].clone(),
+            crate::subagent::provider::resolve_subagent_provider(&settings, &app_config);
-            tool_ctx.clone(),
+
-            "read".to_string(),
+        let base_url = app_config
-            credentials.base_url.clone(),
+            .providers
-            credentials.api_key.clone(),
+            .get(&provider)
-            credentials.model.clone(),
+            .map(|p| p.api_base.clone())
            .unwrap_or_else(|| zesdex_domain::agent::defaults::DEFAULT_API_BASE.to_string());
        let api_key = crate::llm::provider::resolve_api_key(&settings, &app_config);
        let workspace_root = ctx
            .workspaces
            .first()
            .map(|p| p.to_string_lossy().to_string())
            .unwrap_or_else(|| ".".to_string());
        // One lightweight scout — no parallel agents, no index rebuild.
        let directive = format!(
            "{}\n\nUser's task: {goal}\nWorkspace root: {workspace_root}",
            explore_scout_directive()
        );
-        let directive = directives[i].clone();
+        let subagent_ctx = SubagentContext::new(
-        let tc = tool_ctx.clone();
+            directive.clone(),
            ctx.clone(),
            "read".to_string(),
            base_url,
            api_key,
            model,
        );
-        let handle = thread::spawn(move || {
+        let rt = crate::runtime::runtime();
-            let rt =
+        let result = rt.block_on(run_agent(
-                tokio::runtime::Runtime::new().context("create explore subagent tokio runtime")?;
+            subagent_ctx,
-            rt.block_on(run_agent(ctx, &directive, AccessTier::Read, tc))
+            &directive,
-        });
+            AccessTier::Read,
            ctx.clone(),
        ))?;
-        handles.push((i, handle));
+        let mut out = format!("[Codebase scout report]\n{goal}\n\n----------\n{}", result);
-    }
+        if out.len() > EXPLORE_OUTPUT_MAX_CHARS {
-
+            warn!(
-    // ── 4. Join handles via spawn_blocking ──────────────────────────────
+                "explore_codebase output truncated: {} chars -> {}",
-    let turn_events_clone = Arc::clone(turn_events);
+                out.len(),
-    let results: Vec<(usize, String, bool)> = tokio::task::spawn_blocking(move || {
+                EXPLORE_OUTPUT_MAX_CHARS
-        let mut out = Vec::with_capacity(EXPLORE_AGENT_COUNT);
+            );
-        for (i, handle) in handles {
+            out.truncate(EXPLORE_OUTPUT_MAX_CHARS);
-            let entry = match handle.join() {
+            out.push_str("\n...[truncated]");
                Ok(Ok(output)) => {
                    info!(agent = i, "explore subagent completed");
                    emit_completed(&turn_events_clone, EXPLORE_IDS[i], EXPLORE_LABELS[i]);
                    (i, output, true)
                }
                Ok(Err(e)) => {
                    warn!(agent = i, error = %e, "explore subagent failed");
                    emit_failed(
                        &turn_events_clone,
                        EXPLORE_IDS[i],
                        EXPLORE_LABELS[i],
                        &e.to_string(),
                    );
                    (i, format!("Error: {e}"), false)
                }
                Err(e) => {
                    warn!(agent = i, error = ?e, "explore subagent panicked");
                    emit_failed(
                        &turn_events_clone,
                        EXPLORE_IDS[i],
                        EXPLORE_LABELS[i],
                        "thread panicked",
                    );
                    (i, format!("Thread panic: {e:?}"), false)
                }
            };
            out.push(entry);
        }
        out
    })
    .await
    .context("explore join task panicked")?;
    // ── 5. Build consolidated context ───────────────────────────────────
    Ok(build_explore_context(&results))
 }
 // ---------------------------------------------------------------------------
 //  Consolidation
 // ---------------------------------------------------------------------------
 /// Format explore results as a system-level context message.
 fn build_explore_context(results: &[(usize, String, bool)]) -> String {
    let success_count = results.iter().filter(|r| r.2).count();
    let total = results.len();
    let mut msg = format!("[Explore Phase — {success_count}/{total} agents succeeded]\n\n");
    for (i, output, success) in results {
        let label = EXPLORE_LABELS.get(*i).unwrap_or(&"❓ Unknown");
        if *success {
            msg.push_str(&format!("=== {label} ===\n{output}\n\n"));
        } else {
            msg.push_str(&format!("=== {label} (FAILED) ===\n{output}\n\n"));
        }
        Ok(out)
    }
    msg
 }
@@ -34,6 +34,7 @@ pub mod llm;
 pub mod mcp;
 pub mod middleware;
 pub mod persistence;
 pub mod runtime;
 pub mod subagent;
 pub mod tools;
 pub mod utils;
@@ -72,6 +72,55 @@ fn detect_claude_settings_provider() -> Option<(ProviderConfig, Option<String>)>
    ))
 }
 /// Apply a detected Claude provider + custom model onto an `AppConfig`.
 ///
 /// Pure (no I/O) so it can be unit-tested. Flow:
 ///   1. Always `insert`s the "claude" provider (refreshing a possibly stale
 ///      persisted entry with the current base URL + key from settings.json).
 ///   2. Registers known Claude model roles if missing.
 ///   3. Always sets `default_provider = "claude"` and
 ///      `default_model = custom_model.unwrap_or("claude-opus-4-8")` so Opus
 ///      is the default whenever `~/.claude/settings.json` is present.
 fn apply_claude_provider(
    cfg: &mut AppConfig,
    claude_provider: ProviderConfig,
    custom_model: Option<String>,
 ) {
    cfg.providers.insert("claude".to_string(), claude_provider);
    let claude_models: [(&str, &str); 3] = [
        ("claude-opus-4-8", "claude-opus-4-8"),
        ("claude-sonnet-5", "claude-sonnet-5"),
        ("claude-haiku-4-5", "claude-haiku-4-5-20251001"),
    ];
    for (role_name, model_name) in &claude_models {
        cfg.model_roles
            .entry(role_name.to_string())
            .or_insert(ModelRole {
                provider: "claude".to_string(),
                model: model_name.to_string(),
                max_tokens: Some(8192),
                context_window: Some(200_000),
                temperature: Some(0.7),
            });
    }
    if let Some(custom) = &custom_model {
        cfg.model_roles.entry(custom.clone()).or_insert(ModelRole {
            provider: "claude".to_string(),
            model: custom.clone(),
            max_tokens: Some(8192),
            context_window: Some(200_000),
            temperature: Some(0.7),
        });
    }
    // Always prefer the Claude provider + Opus model when settings.json
    // is present — this is the user's explicit custom endpoint choice.
    cfg.default_provider = "claude".to_string();
    cfg.default_model = custom_model.unwrap_or_else(|| "claude-opus-4-8".to_string());
 }
 impl AppConfigRepository for JsonAppConfigRepository {
    fn load(&self, base_dir: &Path) -> Result<AppConfig, RepositoryError> {
        let path = base_dir.join("app_config.json");
@@ -87,41 +136,7 @@ impl AppConfigRepository for JsonAppConfigRepository {
        }
        if let Some((claude_provider, custom_model)) = detect_claude_settings_provider() {
-            cfg.providers
+            apply_claude_provider(&mut cfg, claude_provider, custom_model);
                .entry("claude".to_string())
                .or_insert(claude_provider);
            let claude_models: [(&str, &str); 3] = [
                ("claude-opus-4-8", "claude-opus-4-8"),
                ("claude-sonnet-5", "claude-sonnet-5"),
                ("claude-haiku-4-5", "claude-haiku-4-5-20251001"),
            ];
            for (role_name, model_name) in &claude_models {
                cfg.model_roles
                    .entry(role_name.to_string())
                    .or_insert(ModelRole {
                        provider: "claude".to_string(),
                        model: model_name.to_string(),
                        max_tokens: Some(8192),
                        context_window: Some(200_000),
                        temperature: Some(0.7),
                    });
            }
            if let Some(custom) = &custom_model {
                cfg.model_roles.entry(custom.clone()).or_insert(ModelRole {
                    provider: "claude".to_string(),
                    model: custom.clone(),
                    max_tokens: Some(8192),
                    context_window: Some(200_000),
                    temperature: Some(0.7),
                });
            }
            if cfg.default_provider == defaults.default_provider {
                cfg.default_provider = "claude".to_string();
                cfg.default_model = custom_model.unwrap_or_else(|| "claude-opus-4-8".to_string());
            }
        }
        Ok(cfg)
@@ -134,3 +149,106 @@ impl AppConfigRepository for JsonAppConfigRepository {
        Ok(())
    }
 }
 #[cfg(test)]
 mod tests {
    use super::*;
    use std::collections::HashMap;
    fn claude_provider(base: &str, key: Option<&str>) -> ProviderConfig {
        ProviderConfig {
            api_base: base.to_string(),
            api_key_env: Some("ANTHROPIC_API_KEY".to_string()),
            default_model: Some("claude-opus-4-8".to_string()),
            default_api_key: key.map(|s| s.to_string()),
        }
    }
    #[test]
    fn claude_settings_parse_env() {
        let parsed: ClaudeSettings = serde_json::from_str(
            r#"{"env":{"ANTHROPIC_BASE_URL":"https://9router.example/v1","ANTHROPIC_API_KEY":"sk-test"}}"#,
        )
        .unwrap();
        let env = parsed.env.unwrap();
        assert_eq!(
            env.anthropic_base_url.as_deref(),
            Some("https://9router.example/v1")
        );
        assert_eq!(env.anthropic_api_key.as_deref(), Some("sk-test"));
    }
    #[test]
    fn apply_claude_refreshes_stale_provider_and_sets_opus_default() {
        // Simulate a previously-persisted app_config.json with a STALE claude
        // provider + non-opus default (e.g. user had switched provider).
        let mut cfg = AppConfig {
            providers: {
                let mut m = HashMap::new();
                m.insert(
                    "claude".to_string(),
                    claude_provider("https://old.example/v1", Some("sk-old")),
                );
                m
            },
            model_roles: HashMap::new(),
            default_provider: "router".to_string(),
            default_model: "other-model".to_string(),
            default_context_window: 256_000,
        };
        // Detect returned a fresh provider from ~/.claude/settings.json.
        apply_claude_provider(
            &mut cfg,
            claude_provider("https://9router.example/v1", Some("sk-new")),
            None,
        );
        let claude = cfg.providers.get("claude").unwrap();
        assert_eq!(claude.api_base, "https://9router.example/v1");
        assert_eq!(claude.default_api_key.as_deref(), Some("sk-new"));
        // Insert (not or_insert) → stale entry refreshed.
        assert_eq!(cfg.default_provider, "claude");
        assert_eq!(cfg.default_model, "claude-opus-4-8");
        // Claude model roles registered.
        assert!(cfg.model_roles.contains_key("claude-opus-4-8"));
        assert!(cfg.model_roles.contains_key("claude-sonnet-5"));
        assert!(cfg.model_roles.contains_key("claude-haiku-4-5"));
    }
    #[test]
    fn apply_claude_honors_custom_model_from_settings() {
        let mut cfg = AppConfig::default();
        apply_claude_provider(
            &mut cfg,
            claude_provider("https://9router.example/v1", Some("sk-new")),
            Some("claude-opus-4-8".to_string()),
        );
        assert_eq!(cfg.default_model, "claude-opus-4-8");
        assert!(cfg.model_roles.contains_key("claude-opus-4-8"));
    }
    #[test]
    fn detect_uses_env_creds_as_fallback() {
        // When ~/.claude/settings.json is absent/unreadable, the env-var
        // fallback should produce a "claude" provider. Set env vars, call
        // detect, and assert the resulting provider uses them.
        std::env::set_var("ANTHROPIC_BASE_URL", "https://env.example/v1");
        std::env::set_var("ANTHROPIC_API_KEY", "sk-env");
        match detect_claude_settings_provider() {
            Some((provider, _custom)) => {
                // If the real settings.json exists it wins (base could be the
                // real 9router URL); otherwise env creds are used. Either way,
                // the provider must have api_key_env pointing at ANTHROPIC_API_KEY.
                assert_eq!(provider.api_key_env.as_deref(), Some("ANTHROPIC_API_KEY"));
            }
            None => {
                // No file + no env (shouldn't happen since we just set env).
                panic!("expected env fallback to produce a provider");
            }
        }
        std::env::remove_var("ANTHROPIC_BASE_URL");
        std::env::remove_var("ANTHROPIC_API_KEY");
    }
 }
@@ -0,0 +1,62 @@
 //! Process-wide shared Tokio runtime for sync → async bridging.
 //!
 //! Many `Tool::run` implementations are synchronous but need to drive async
 //! work (LLM calls, subagent execution). Creating a fresh
 //! [`tokio::runtime::Runtime`] on every call is expensive (spawns a thread
 //! pool + runtime each time) and can fail randomly under thread pressure.
 //!
 //! # Flow
 //!
 //! [`runtime()`] returns a lazily-initialised process-wide runtime created
 //! exactly once via [`std::sync::OnceLock`]. Callers use
 //! `runtime().block_on(...)` exactly like they would with a local runtime —
 //! the only difference is the runtime is shared, so the cost is paid once per
 //! process instead of once per tool call.
 //!
 //! # Safety
 //!
 //! `block_on` panics if called from within a running Tokio runtime. The
 //! tools that use this helper are synchronous (`Tool::run`), so this is safe
 //! in practice. Async code should never call `runtime().block_on`.
 use std::sync::OnceLock;
 /// Maximum worker threads for the shared runtime. Kept modest — tools are
 /// mostly I/O-bound and rarely need more concurrency than this.
 const RUNTIME_WORKER_THREADS: usize = 8;
 static SHARED_RUNTIME: OnceLock<tokio::runtime::Runtime> = OnceLock::new();
 /// Return the process-wide shared Tokio runtime, initialising it on first use.
 ///
 /// The runtime is configured with `worker_threads = 8` and
 /// `enable_all()` (time + IO drivers) so streams, timers, and network calls
 /// all work. If initialisation fails (extremely rare — resource exhaustion at
 /// startup), the process aborts with a clear message rather than returning
 /// an error on every subsequent call.
 pub fn runtime() -> &'static tokio::runtime::Runtime {
    SHARED_RUNTIME.get_or_init(|| {
        tokio::runtime::Builder::new_multi_thread()
            .worker_threads(RUNTIME_WORKER_THREADS)
            .thread_name("zesdex-shared-rt")
            .enable_all()
            .build()
            .expect("failed to create shared tokio runtime")
    })
 }
 #[cfg(test)]
 mod tests {
    use super::runtime;
    #[test]
    fn runtime_is_singleton() {
        assert!(std::ptr::eq(runtime(), runtime()));
    }
    #[test]
    fn runtime_blocks_and_resolves() {
        let val = runtime().block_on(async { 6 * 7 });
        assert_eq!(val, 42);
    }
 }
@@ -23,6 +23,37 @@ use zesdex_domain::subagent_directive;
 /// Maximum number of tool-call iterations before the engine gives up.
 const MAX_ITERATIONS: u32 = 25;
 /// A single tool-result message is truncated before entering the subagent's
 /// context so it cannot blow the window (matches the main turn service).
 const TOOL_OUTPUT_MAX_CHARS: usize = 12_000;
 /// Maximum consecutive identical tool errors before the engine injects a
 /// recovery note steering the model to a different approach.
 const MAX_CONSECUTIVE_TOOL_ERRORS: usize = 3;
 /// Pick a `max_tokens` budget proportional to the directive's length.
 fn adaptive_max_tokens(directive_len: usize) -> u32 {
    if directive_len <= 80 {
        800
    } else if directive_len <= 400 {
        1600
    } else {
        4096
    }
 }
 fn truncate_tool_output(output: String) -> String {
    if output.len() <= TOOL_OUTPUT_MAX_CHARS {
        return output;
    }
    let mut result: String = output.chars().take(TOOL_OUTPUT_MAX_CHARS).collect();
    result.push_str(&format!(
        "\n...[truncated {} chars]",
        output.len() - TOOL_OUTPUT_MAX_CHARS
    ));
    result
 }
 /// Emit an `AgentProgress` event onto the turn-event queue, if one is
 /// configured in the `ToolCtx`.
 fn report_progress(tool_ctx: &ToolCtx, progress: AgentProgress) {
@@ -85,11 +116,17 @@ pub async fn run_agent(
        Some(ctx.base_url.clone()),
    );
    let max_tokens = adaptive_max_tokens(directive.len());
    // Track repeated tool errors so the agent can recover from a dead end.
    let mut consecutive_errors = 0usize;
    let mut last_tool = String::new();
    // Limited iteration loop so we don't run forever
    for iteration in 0..MAX_ITERATIONS {
        use zesdex_application::ports::ProviderService;
        let (response_msg, _usage) = client
-            .chat(&messages, Some(defs.clone()), Some(4096), None)
+            .chat(&messages, Some(defs.clone()), Some(max_tokens), Some(0.2))
            .await?;
        let content = response_msg.content.clone().unwrap_or_default();
@@ -113,7 +150,7 @@ pub async fn run_agent(
                &tool_ctx,
                AgentProgress::running(
                    "subagent",
-                    format!("{}:{}", directive, tool_name),
+                    format!("{}:{tool_name}", directive),
                    Some(tool_name.clone()),
                ),
            );
@@ -127,7 +164,29 @@ pub async fn run_agent(
                format!("Unknown tool: {tool_name}")
            };
-            messages.push(ChatMessage::tool(tc.id.clone(), result));
+            // Error-recovery: if the same tool keeps failing, inject a
            // system note steering the model to a different approach.
            if result.starts_with("Error:") {
                if last_tool.as_str() == tool_name.as_str() {
                    consecutive_errors += 1;
                } else {
                    consecutive_errors = 1;
                    last_tool = tool_name.to_string();
                }
                if consecutive_errors >= MAX_CONSECUTIVE_TOOL_ERRORS {
                    messages.push(ChatMessage::system(
                        zesdex_domain::agent::prompt::error_recovery_note(tool_name, &result),
                    ));
                    consecutive_errors = 0;
                }
            } else {
                consecutive_errors = 0;
            }
            messages.push(ChatMessage::tool(
                tc.id.clone(),
                truncate_tool_output(result),
            ));
        }
        // Add assistant response if there was text content
@@ -63,15 +63,21 @@ impl SubagentProvider {
 /// Resolve subagent provider and model from settings.
 ///
-/// Flow: reads `settings.provider` and `settings.model` → if model is empty,
+/// Flow: reads `settings.provider` and `settings.model` → if provider is
 /// empty, falls back to `app_config.default_provider` → if model is empty,
 /// falls back to the provider config's `default_model` → if that is also
-/// empty, uses the domain default model constant.
+/// empty, uses `app_config.default_model` → finally the domain default model
 /// constant.
 #[instrument]
 pub fn resolve_subagent_provider(
    settings: &zesdex_domain::cms::Settings,
    app_config: &zesdex_domain::cms::AppConfig,
 ) -> (String, String) {
-    let provider = settings.provider.clone();
+    let provider = if settings.provider.is_empty() {
        app_config.default_provider.clone()
    } else {
        settings.provider.clone()
    };
    let model = settings.model.clone();
    // Use the default model from the provider config if available
@@ -80,6 +86,7 @@ pub fn resolve_subagent_provider(
            .providers
            .get(&provider)
            .and_then(|p| p.default_model.clone())
            .or_else(|| Some(app_config.default_model.clone()))
            .unwrap_or_else(|| zesdex_domain::agent::defaults::DEFAULT_MODEL.to_string())
    } else {
        model
@@ -32,7 +32,6 @@ pub fn spawn_subagent(
 ) -> thread::JoinHandle<Result<String>> {
    info!("Spawning subagent: {directive}");
    thread::spawn(move || {
-        let rt = tokio::runtime::Runtime::new()?;
+        crate::runtime::runtime().block_on(run_agent(ctx, &directive, access, tool_ctx))
        rt.block_on(run_agent(ctx, &directive, access, tool_ctx))
    })
 }
@@ -123,7 +123,7 @@ impl Tool for ParallelDelegate {
                .collect()
        } else {
            // Auto-split using LLM
-            let rt = tokio::runtime::Runtime::new()?;
+            let rt = crate::runtime::runtime();
            let directives = rt.block_on(auto_split_task(
                &task,
                max_parallel,
@@ -146,49 +146,53 @@ impl Tool for ParallelDelegate {
            "parallel delegation: starting subagents"
        );
-        // Spawn agents in parallel
+        // Spawn agents in parallel — bounded: never more than `max_parallel`
-        let mut handles = Vec::new();
+        // subagent threads in flight at once (Claude Code-style isolation).
        for (i, (directive, access)) in directives.iter().enumerate() {
            let subagent_ctx = SubagentContext::new(
                directive.clone(),
                ctx.clone(),
                format!("{access:?}"),
                base_url.clone(),
                api_key.clone(),
                model.clone(),
            );
            debug!(agent_index = i, access = ?access, "spawning parallel agent");
            let handle = spawn_subagent(subagent_ctx, directive.clone(), *access, ctx.clone());
            handles.push((i, handle));
        }
        // Join all results
        let mut results: Vec<(usize, String, String)> = Vec::new();
-        for (i, handle) in handles {
+        for batch in directives.chunks(max_parallel) {
-            match handle.join() {
+            let mut handles = Vec::with_capacity(batch.len());
-                Ok(Ok(output)) => {
+            for (i, (directive, access)) in batch.iter().enumerate() {
-                    info!(agent_index = i, "parallel agent completed");
+                let global_idx = results.len() + i;
-                    results.push((i, directives[i].0.clone(), output));
+                let subagent_ctx = SubagentContext::new(
-                }
+                    directive.clone(),
-                Ok(Err(e)) => {
+                    ctx.clone(),
-                    warn!(agent_index = i, error = %e, "parallel agent failed");
+                    format!("{access:?}"),
-                    results.push((i, directives[i].0.clone(), format!("[ERROR] {e}")));
+                    base_url.clone(),
-                }
+                    api_key.clone(),
-                Err(e) => {
+                    model.clone(),
-                    warn!(agent_index = i, error = ?e, "parallel agent panicked");
+                );
-                    results.push((
+
-                        i,
+                debug!(agent_index = global_idx, access = ?access, "spawning parallel agent");
-                        directives[i].0.clone(),
+                let handle = spawn_subagent(subagent_ctx, directive.clone(), *access, ctx.clone());
-                        "[ERROR] Agent panicked".to_string(),
+                handles.push((global_idx, handle));
-                    ));
+            }
            // Join this batch before spawning the next.
            for (i, handle) in handles {
                match handle.join() {
                    Ok(Ok(output)) => {
                        info!(agent_index = i, "parallel agent completed");
                        results.push((i, directives[i].0.clone(), output));
                    }
                    Ok(Err(e)) => {
                        warn!(agent_index = i, error = %e, "parallel agent failed");
                        results.push((i, directives[i].0.clone(), format!("[ERROR] {e}")));
                    }
                    Err(e) => {
                        warn!(agent_index = i, error = ?e, "parallel agent panicked");
                        results.push((
                            i,
                            directives[i].0.clone(),
                            "[ERROR] Agent panicked".to_string(),
                        ));
                    }
                }
            }
        }
        // Consolidate results
        if synthesize && results.len() > 1 {
-            let rt = tokio::runtime::Runtime::new()?;
+            let rt = crate::runtime::runtime();
            let consolidated =
                rt.block_on(consolidate_results(&results, &base_url, &api_key, &model))?;
            Ok(format!(
@@ -42,6 +42,7 @@ pub fn all_tools() -> Vec<Box<dyn super::Tool>> {
        // ── Best-practice tools (built-in) ─────────────────────────
        Box::new(super::best_practice::BestPractice),
        Box::new(super::best_practice::CommitConvention),
        Box::new(crate::best_practice::explore::ExploreCodebase),
    ]
 }
@@ -63,7 +63,7 @@ impl crate::tools::Tool for DirCacheUpdate {
        // Persist the resolved paths into the shared DirCache so the TUI
        // and other tools can read the cached listing without re-scanning.
        let dc = ctx.dir_cache.clone();
-        let rt = tokio::runtime::Runtime::new()?;
+        let rt = crate::runtime::runtime();
        rt.block_on(async { dc.write().await.set(resolved).await });
        info!(count, "directory cache updated");
@@ -65,7 +65,7 @@ impl Tool for WorkflowRun {
            zesdex_domain::agent::defaults::DEFAULT_MODEL.to_string(),
            None,
        );
-        let rt = tokio::runtime::Runtime::new()?;
+        let rt = crate::runtime::runtime();
        let result: Vec<String> =
            rt.block_on(async { execute_workflow(&script, ctx, &llm_client).await })?;
@@ -229,7 +229,7 @@ impl Tool for HiveMind {
            .ok_or_else(|| anyhow::anyhow!("missing 'cycles' array"))?;
        info!("Hive mind starting with {} cycles", cycles_val.len());
-        let rt = tokio::runtime::Runtime::new()?;
+        let rt = crate::runtime::runtime();
        let mut all_node_outputs = Vec::new();
        for (cycle_idx, cycle_val) in cycles_val.iter().enumerate() {
@@ -1,11 +1,15 @@
 //! Hive-mind cycle execution — run one cycle of parallel nodes.
 //!
 //! Flow: load settings → resolve LLM credentials → run all directives in the
-//! cycle concurrently via try_join_all → collect Vec<NodeOutput>.
+//! cycle concurrently via a BOUNDED buffer (`buffer_unordered(MAX)`) → collect
 //! `Vec<NodeOutput>`. Unlike `try_join_all`, a single failing node does NOT
 //! fail the whole cycle — failed nodes are logged and replaced with an
 //! `[ERROR]` output so the remaining results are preserved (like Claude
 //! Code's isolated subagents).
 use anyhow::Result;
-use futures_util::future::try_join_all;
+use futures_util::stream::StreamExt;
-use tracing::info;
+use tracing::{info, warn};
 use zesdex_domain::cms::{AppConfigRepository, SettingsRepository};
 use zesdex_domain::core::Store;
@@ -15,16 +19,21 @@ use crate::subagent::context::SubagentContext;
 use crate::subagent::division::AccessTier;
 use crate::subagent::engine::run_agent;
 use crate::tools::ToolCtx;
-use zesdex_domain::workflow::{CognitiveCycle, NodeOutput};
+use zesdex_domain::workflow::{CognitiveCycle, NodeDirective, NodeOutput};
 /// Maximum number of hive-mind nodes running concurrently per cycle.
 /// Keeps thread/runtime pressure bounded (Claude Code-style).
 const MAX_CONCURRENT_NODES: usize = 8;
 /// Execute one cycle: run each node directive and collect outputs.
 ///
 /// Flow:
 ///   1. Load `Settings` and `AppConfig` from the store directory.
 ///   2. Resolve provider, model, base_url, and api_key.
-///   3. Spawn all directives concurrently — each builds a `SubagentContext`
+///   3. Spawn directives with bounded concurrency — each builds a
-///      and calls `run_agent` (Full access).
+///      `SubagentContext` and calls `run_agent`.
-///   4. `try_join_all` waits for all to complete, then collect `NodeOutput`s.
+///   4. Collect `NodeOutput`s; failed nodes are logged and replaced with an
 ///      `[ERROR]` placeholder so the cycle still completes.
 pub async fn execute_cycle(cycle: &CognitiveCycle, tool_ctx: &ToolCtx) -> Result<Vec<NodeOutput>> {
    info!(
        "Executing cycle {} with {} directives",
@@ -53,10 +62,9 @@ pub async fn execute_cycle(cycle: &CognitiveCycle, tool_ctx: &ToolCtx) -> Result
    let cycle_index = cycle.index;
-    use zesdex_domain::workflow::NodeDirective;
+    // Run all directives with bounded concurrency. Each node is its own
-
+    // future; failures are collected, not propagated (isolated errors).
-    // Run all directives in this cycle concurrently.
+    let tasks: Vec<_> = cycle
    let handles: Vec<_> = cycle
        .directives
        .iter()
        .enumerate()
@@ -79,18 +87,33 @@ pub async fn execute_cycle(cycle: &CognitiveCycle, tool_ctx: &ToolCtx) -> Result
                _ => AccessTier::Read,
            };
            let node_id = format!("Node-{}-{}", cycle_index, i);
            async move {
-                let result = run_agent(ctx, &dir, access, tc).await?;
+                match run_agent(ctx, &dir, access, tc).await {
-                Ok::<NodeOutput, anyhow::Error>(NodeOutput {
+                    Ok(output) => Ok::<NodeOutput, anyhow::Error>(NodeOutput {
-                    id: format!("Node-{}-{}", cycle_index, i),
+                        id: node_id.clone(),
-                    directive: dir,
+                        directive: dir,
-                    output: result,
+                        output,
-                })
+                    }),
                    Err(e) => {
                        warn!(node = %node_id, error = %e, "hive-mind node failed (isolated)");
                        Ok::<NodeOutput, anyhow::Error>(NodeOutput {
                            id: node_id,
                            directive: dir,
                            output: format!("[ERROR] {e}"),
                        })
                    }
                }
            }
        })
        .collect();
-    let results = try_join_all(handles).await?;
+    // Bounded concurrency: run at most MAX_CONCURRENT_NODES futures at once.
    let mut stream = futures_util::stream::iter(tasks).buffer_unordered(MAX_CONCURRENT_NODES);
    let mut results = Vec::with_capacity(cycle.directives.len());
    while let Some(node) = stream.next().await {
        results.push(node?);
    }
    Ok(results)
 }
@@ -471,8 +471,7 @@ fn handle_compact(state: &mut AppStateRest) {
    let client = zesdex_infrastructure::llm::provider::LlmClient::new(api_key, model, api_base);
    if let Some(ref mut rt) = state.session_runtime {
-        let tokio_rt =
+        let tokio_rt = zesdex_infrastructure::runtime::runtime();
            tokio::runtime::Runtime::new().expect("create tokio runtime for AI compaction");
        if let Ok(()) = tokio_rt.block_on(
            zesdex_application::agent::turn_service::compact_messages_with_ai(
                &mut rt.messages,
@@ -110,13 +110,6 @@ pub fn spawn_agent_turn(state: &mut AppStateRest, text: String) {
        api_base: api_base.clone(),
    };
    // Clone credentials before moving into LlmClient.
    let explore_api_key = api_key.clone();
    let explore_model = model.clone();
    let explore_base_url = api_base
        .clone()
        .unwrap_or_else(|| "https://api.openai.com/v1".to_string());
    let client = std::sync::Arc::new(LlmClient::new(api_key, model, api_base));
    let tool_ctx = ToolCtx::builder()
@@ -125,29 +118,12 @@ pub fn spawn_agent_turn(state: &mut AppStateRest, text: String) {
        .turn_events(turn_events)
        .build();
    // Clone ToolCtx for the explore service (before moving into executor).
    let explore_ctx = tool_ctx.clone();
    let tool_executor = std::sync::Arc::new(InfrastructureToolExecutor::new(tool_ctx));
    let tools = all_tools();
    let defs = tool_defs(&tools);
-    // Wire the mandatory explore phase (3+ parallel subagents).
+    let turn_service = AgentTurnServiceImpl::new(client, tool_executor, defs);
    let explore_creds = zesdex_infrastructure::best_practice::explore::Credentials {
        base_url: explore_base_url,
        api_key: explore_api_key,
        model: explore_model,
    };
    let explore_service = std::sync::Arc::new(
        zesdex_infrastructure::best_practice::explore::ExploreServiceImpl::new(
            explore_ctx,
            explore_creds,
        ),
    );
    let turn_service =
        AgentTurnServiceImpl::new(client, tool_executor, defs).with_explore(explore_service);
    tokio::spawn(async move {
        let _ = turn_service.run_turn(params).await;
Author	SHA1	Message	Date
semantic-release-bot	4dccf0cee4	chore(release): 1.19.3 [skip ci] ## [1.19.3](https://github.com/asepharyana/zesdex/compare/v1.19.2...v1.19.3) (2026-08-27) ### Bug Fixes * api: model Opus pakai URL + API custom dari ~/.claude/settings.json ([`1f91b44`](https://github.com/asepharyana/zesdex/commit/1f91b447080e7106201d77585edb7d38b74e34bb))	2026-08-27 16:49:19 +00:00
asepharyana	1f91b44708	fix(api): model Opus pakai URL + API custom dari ~/.claude/settings.json Perbaiki provider claude agar selalu refresh dari settings.json dan menjadi default (claude-opus-4-8) setiap startup: - fix(app_config_repo): ganti or_insert -> insert untuk provider claude — base_url/key dari ~/.claude/settings.json selalu di-refresh, tidak tertutup snapshot lama app_config.json. - fix(app_config_repo): hapus kondisi default_provider == default — saat settings.json terdeteksi, default_provider='claude' dan default_model='claude-opus-4-8' SELALU di-set (sebelumnya skip kalau user pernah ganti provider). - fix(subagent/provider): resolve_subagent_provider fallback ke app_config.default_provider/default_model kalau settings.provider/model kosong — subagent ikut pakai Opus. - test: 4 unit test (parse settings.json, refresh stale provider, custom model, env fallback). Verified live: settings.json terbaca (9router URL + key).	2026-08-27 23:45:30 +07:00
semantic-release-bot	b25929824a	chore(release): 1.19.2 [skip ci] ## [1.19.2](https://github.com/asepharyana/zesdex/compare/v1.19.1...v1.19.2) (2026-08-27) ### Performance Improvements * agent: stabilkan async & parallel — satu runtime, bounded concurrency, isolasi error ([`6a98d52`](https://github.com/asepharyana/zesdex/commit/6a98d52d54a69f78710d852a69dda3ac0a4ead31))	2026-08-27 16:30:31 +00:00
asepharyana	3fd9a2b2db	chore: sinkronkan Cargo.lock dengan versi 1.19.1	2026-08-27 23:26:37 +07:00
asepharyana	6a98d52d54	perf(agent): stabilkan async & parallel — satu runtime, bounded concurrency, isolasi error Seperti Claude Code: satu runtime shared, concurrency dibatasi, error subagent terisolasi (satu node gagal tidak menggagalkan cycle). - feat(runtime): global tokio runtime via OnceLock — ganti 9+ titik Runtime::new() per tool call (spawn, parallel_delegate, workflow, explore, dir_cache, daemon handler). Hemat resource, hilangkan panic path Runtime::new().expect() di daemon compaction. - fix(workflow): execute_cycle ganti try_join_all (fail-fast) → buffer_unordered(8) + isolasi error per node; node gagal di-log dan diganti [ERROR], hasil node lain tetap dipakai (Claude Code-style). - fix(parallel_delegate): spawn subagent dibatasi per batch max_parallel (tidak unbounded threads). - perf(subagent): run_agent adaptif max_tokens (800/1600/4096), temp 0.2, truncate tool output 12k, error-recovery note utk tool error berulang. - test: runtime singleton + block_on (2 test).	2026-08-27 23:25:44 +07:00
semantic-release-bot	3847c0e6fd	chore(release): 1.19.1 [skip ci] ## [1.19.1](https://github.com/asepharyana/zesdex/compare/v1.19.0...v1.19.1) (2026-08-27) ### Performance Improvements * agent: rombak alur AI agent — adaptif, hemat token, self-healing ([`eac0443`](https://github.com/asepharyana/zesdex/commit/eac0443c4c3b8bfcbefd4bad9554168fb6525b94))	2026-08-27 16:10:48 +00:00
asepharyana	5023e5dfa1	chore: sinkronkan Cargo.lock dengan versi 1.19.0	2026-08-27 23:06:57 +07:00
asepharyana	eac0443c4c	perf(agent): rombak alur AI agent — adaptif, hemat token, self-healing Ganti explore phase MANDATORY (3 subagent tiap turn, boros) dengan tool explore_codebase yang DIPUTUSKAN agent sendiri (lazy, token-aware): - hapus ExploreService trait + with_explore + Phase 0 dari turn loop - ExploreServiceImpl kini jadi tool 'explore_codebase' (1 context-scout subagent, read-only, cap output 4k chars) - system prompt: instruksi TOKEN BUDGET (jawab langsung utk query simple, panggil explore_codebase sekali utk task kompleks) Loop utama kini adaptif & self-healing: - max_tokens adaptif (800/1600/4096 by request length) — bukan selalu 4096 - temperature 0.2 saat tool-calling, 0.7 utk final answer - ErrorTracker: deteksi tool error berulang → inject recovery note, stop setelah 8 error total (bukan 50 iterasi sia-sia) - auto-compact history > 60k chars sebelum LLM call - tool output di-truncate ke 12k chars sebelum masuk konteks Tambah 8 unit test (truncation, adaptive tokens, error tracker).	2026-08-27 23:06:37 +07:00
asepharyana	14f3eae62a	a	2026-08-27 23:06:37 +07:00