deepagents-code 0.1.36__tar.gz → 0.1.37__tar.gz

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/AGENTS.md

@@ -91,7 +91,8 @@ Debug logging is configured **once**, on the `deepagents_code` package logger, b
 
 - Do **not** add per-module `configure_debug_logging(logger)` calls. They are redundant now that the package logger is configured at import, and they reintroduce the duplicate-handler problem the single-config approach solves.
 - Every module should create its logger with `logging.getLogger(__name__)` so it stays inside the `deepagents_code.*` hierarchy and inherits the package handler. Don't set `logger.propagate = False` or attach your own handlers.
-- The handler only attaches when `DEEPAGENTS_CODE_DEBUG` is truthy; the no-op path is a single env-var read, so it's safe on the startup hot path. See `DEVELOPMENT.md` for the `DEEPAGENTS_CODE_DEBUG` / `DEEPAGENTS_CODE_DEBUG_FILE` env vars.
+- The **file** handler only attaches when `DEEPAGENTS_CODE_DEBUG` is truthy; that path is a single env-var read, so it's safe on the startup hot path. See `DEVELOPMENT.md` for the `DEEPAGENTS_CODE_DEBUG` / `DEEPAGENTS_CODE_DEBUG_FILE` / `DEEPAGENTS_CODE_LOG_LEVEL` env vars.
+- Separately, an **always-on in-memory ring buffer** (`_debug_buffer.install_log_buffer`, called from `__init__.py` right before `configure_debug_logging`) attaches unconditionally at import so the in-app Debug Console (`Ctrl+\`) can tail recent `deepagents_code.*` records without file logging. It captures `INFO` and above by default (or `DEEPAGENTS_CODE_LOG_LEVEL`), and may lower the package logger to `INFO` for the process lifetime. It never spills to the terminal (a handler is always found, so `lastResort` is skipped) and is bounded (a `deque` of 1000 records), so it's cheap enough to keep on. Because it installs *before* `configure_debug_logging`, warnings that helper emits at startup are captured and visible in the console.
 
 ## CLI help screen
 

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/CHANGELOG.md

@@ -2,6 +2,46 @@
 
 # Deep Agents Code Changelog
 
+## [0.1.37](https://github.com/langchain-ai/deepagents/compare/deepagents-code==0.1.36...deepagents-code==0.1.37) (2026-07-13)
+
+### Features
+
+* Add Meta model provider ([#4650](https://github.com/langchain-ai/deepagents/issues/4650)) ([70829c5](https://github.com/langchain-ai/deepagents/commit/70829c5846b6bdde1cee51f0f4929e819ba1026b))
+* Set `prompt_cache_key` for OpenAI models ([#4632](https://github.com/langchain-ai/deepagents/issues/4632)) ([8cf57ac](https://github.com/langchain-ai/deepagents/commit/8cf57aca9f41c3c4ca6d9796dc2cb80d4b22fb6d))
+* Support Fireworks `/routers` model ids ([#4591](https://github.com/langchain-ai/deepagents/issues/4591)) ([1c08d27](https://github.com/langchain-ai/deepagents/commit/1c08d2705f73f6870610849e00910b94458accf1))
+* `/model` Ctrl+N toggle for names vs raw specs ([#4592](https://github.com/langchain-ai/deepagents/issues/4592)) ([518c322](https://github.com/langchain-ai/deepagents/commit/518c322e7dda840b5a740541afc895516a74925f))
+* `/tools` slash command ([#4649](https://github.com/langchain-ai/deepagents/issues/4649)) ([b1600a8](https://github.com/langchain-ai/deepagents/commit/b1600a8da784899d9db5bdf93be07b4c7f53a46c))
+* Add `-s` alias for `--skill` ([#4620](https://github.com/langchain-ai/deepagents/issues/4620)) ([c9b7ac2](https://github.com/langchain-ai/deepagents/commit/c9b7ac20752e4270d286b4af3fcf710e7e5bee87))
+* Configurable chat cursor style ([#4687](https://github.com/langchain-ai/deepagents/issues/4687)) ([a22484b](https://github.com/langchain-ai/deepagents/commit/a22484ba2e001fd62fbff0fe916f1b3b93889fc3))
+* Expand environment variables in MCP config ([#4681](https://github.com/langchain-ai/deepagents/issues/4681)) ([4f5d7be](https://github.com/langchain-ai/deepagents/commit/4f5d7be1b4073dee346e5a974841721646fa4860))
+* Hide diff widget for credential files ([#4593](https://github.com/langchain-ai/deepagents/issues/4593)) ([4c49a24](https://github.com/langchain-ai/deepagents/commit/4c49a24c4c488f45859b45628cfa57719e52f596))
+* In-app Debug Console ([#4564](https://github.com/langchain-ai/deepagents/issues/4564)) ([4f94a30](https://github.com/langchain-ai/deepagents/commit/4f94a30c11efb1819b08647d15b50b9b1e7bf043))
+* Offer abort in `/threads` cwd-switch prompt ([#4583](https://github.com/langchain-ai/deepagents/issues/4583)) ([aaeac99](https://github.com/langchain-ai/deepagents/commit/aaeac99d9f17b6a55d663c21a4d063444b6a1870))
+* Resume threads in-TUI with `/threads -r [ID]` ([#4609](https://github.com/langchain-ai/deepagents/issues/4609)) ([d442673](https://github.com/langchain-ai/deepagents/commit/d44267358e272553c8ece33d367514430be7c0da))
+* Show `(debug enabled)` on splash when `DEEPAGENTS_CODE_DEBUG` is set ([#4584](https://github.com/langchain-ai/deepagents/issues/4584)) ([f10b877](https://github.com/langchain-ai/deepagents/commit/f10b877e5e48c3a1e15bd72d2ce06dfb16203231))
+
+### Bug Fixes
+
+* Avoid repeated startup auto-update stalls ([#4648](https://github.com/langchain-ai/deepagents/issues/4648)) ([12a9c9d](https://github.com/langchain-ai/deepagents/commit/12a9c9d6813a83e5eeff9feed0c4068d1fcc69c4))
+* Infer Fireworks provider from qualified model IDs ([#4594](https://github.com/langchain-ai/deepagents/issues/4594)) ([4d2aa8a](https://github.com/langchain-ai/deepagents/commit/4d2aa8a9684609eee45cd93b0af5149d8dd09eea))
+* Capture input typed before TUI startup ([#4684](https://github.com/langchain-ai/deepagents/issues/4684)) ([ef9a4a8](https://github.com/langchain-ai/deepagents/commit/ef9a4a8770f4daec70e640d6d2f23dd798d4ede2))
+* Detach owned `langgraph dev` server from terminal ([#4642](https://github.com/langchain-ai/deepagents/issues/4642)) ([d1f3afe](https://github.com/langchain-ai/deepagents/commit/d1f3afecdc015b5d1c9fb098d6347c7d7180e417))
+* Infer additional model providers ([#4675](https://github.com/langchain-ai/deepagents/issues/4675)) ([4ceed24](https://github.com/langchain-ai/deepagents/commit/4ceed24d3758801ed9df1024602e90488004cef1))
+* Preserve `Ctrl+D` deletion in non-empty input ([#4626](https://github.com/langchain-ai/deepagents/issues/4626)) ([306bd89](https://github.com/langchain-ai/deepagents/commit/306bd893f214237b6b4eede7c2ce0eaff66f5527))
+* Quit with `Ctrl+D` at end of prompt ([#4678](https://github.com/langchain-ai/deepagents/issues/4678)) ([2f8c111](https://github.com/langchain-ai/deepagents/commit/2f8c11176f5cc9350c3337b2dfcf9f19886d5d02))
+* Reap langgraph dev server when startup is cancelled ([#4629](https://github.com/langchain-ai/deepagents/issues/4629)) ([904ff05](https://github.com/langchain-ai/deepagents/commit/904ff05620a576100e05fd65810d38405c4942d4))
+* Reject `--auto-approve` in headless mode ([#4617](https://github.com/langchain-ai/deepagents/issues/4617)) ([997be16](https://github.com/langchain-ai/deepagents/commit/997be1643aff6d6900b8b8c95832cdf93be8c0d7))
+* Route explicit `--stdin` + `--skill` to headless path ([#4611](https://github.com/langchain-ai/deepagents/issues/4611)) ([724e24a](https://github.com/langchain-ai/deepagents/commit/724e24a31546e1f31feca3bb2344717550682595))
+* Skip Esc prompt-restore once output generation begins ([#4582](https://github.com/langchain-ai/deepagents/issues/4582)) ([14f384f](https://github.com/langchain-ai/deepagents/commit/14f384fc0083c07a7f44f97543b40b74cf93c13f))
+* Remove misleading agent names from help ([#4671](https://github.com/langchain-ai/deepagents/issues/4671)) ([ac15732](https://github.com/langchain-ai/deepagents/commit/ac1573281530c1564fe07f8a21608e4d4dec2c3b))
+* Support plain `exit` quit command ([#4543](https://github.com/langchain-ai/deepagents/issues/4543)) ([e6f10a1](https://github.com/langchain-ai/deepagents/commit/e6f10a149051128677c9f6f37663922054a7ac26))
+* Sync `ask_user` active-question highlight with focus ([#4599](https://github.com/langchain-ai/deepagents/issues/4599)) ([e4c29b5](https://github.com/langchain-ai/deepagents/commit/e4c29b5264413d23d8dc70cabae72feb5f3f3dcb))
+* Wrap MCP viewer navigation ([#4677](https://github.com/langchain-ai/deepagents/issues/4677)) ([cffc732](https://github.com/langchain-ai/deepagents/commit/cffc73213bfdcec26cb82c413fa7e27db4afc0ca))
+
+### Performance Improvements
+
+* Load MCP servers concurrently during graph build ([#4659](https://github.com/langchain-ai/deepagents/issues/4659)) ([c5345cc](https://github.com/langchain-ai/deepagents/commit/c5345cc04cd810c6238e42dfbfa14497ae1b9020))
+
 ## [0.1.36](https://github.com/langchain-ai/deepagents/compare/deepagents-code==0.1.35...deepagents-code==0.1.36) (2026-07-09)
 
 ### Features

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/COMMANDS.md

@@ -8,7 +8,7 @@ Regenerate this file with `make commands-catalog` after changing command names,
 aliases, descriptions, visibility, or hidden-command metadata.
 
 
-## Public (32)
+## Public (33)
 
 | Command | Aliases | Description |
 | --- | --- | --- |
@@ -41,12 +41,14 @@ aliases, descriptions, visibility, or hidden-command metadata.
 | `/threads` |  | Browse and resume past threads |
 | `/timestamps` |  | Show or hide message timestamps |
 | `/tokens` |  | Show token usage |
+| `/tools` |  | List the tools available to the agent |
 | `/trace` |  | Open this thread in LangSmith |
 | `/update` |  | Check for and install updates |
 | `/version` | `/about` | Show version information |
 
-## Hidden (1)
+## Hidden (2)
 
 These commands are intentionally omitted from autocomplete and help. See the `HIDDEN_COMMANDS` docstring in the registry for context.
 
+- `/debug`
 - `/debug-error`

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/DEVELOPMENT.md

@@ -117,6 +117,12 @@ tail -f /tmp/deepagents_debug.log
 
 To send it elsewhere, also `export DEEPAGENTS_CODE_DEBUG_FILE=<path>`. The handler appends across runs, so a single file accumulates every session.
 
+### In-app Debug Console (`Ctrl+\`)
+
+Press `Ctrl+\` (or run the hidden `/debug` command) inside a session to toggle a read-only Debug Console overlay. It shows a point-in-time session/runtime snapshot (version, model, thread, cwd, auto-approve, sandbox, MCP servers, token usage, debug-log path) plus a live tail of recent `deepagents_code.*` log records.
+
+The tail is fed by an always-on in-memory ring buffer (`_debug_buffer.install_log_buffer`, installed on the package logger in `__init__.py`), so it works **without** `DEEPAGENTS_CODE_DEBUG` — though enabling that switch raises the captured level to `DEBUG` and adds the file handler above. In the console, `Ctrl+L` clears the on-screen view (the buffer keeps accruing) and `c` copies the visible lines.
+
 ## Local dev installs
 
 A *local dev install* gives you a persistent `dcode-dev` command that launches your checkout directly. It lives in a dedicated editable venv under `~/.local/share/dcode-dev`, symlinked into `~/.local/bin/dcode-dev`. It can sit alongside a released `dcode` without interfering:

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: deepagents-code
-Version: 0.1.36
-Summary: Terminal interface for Deep Agents - interactive AI agent with file operations, shell access, and sub-agent capabilities.
+Version: 0.1.37
+Summary: Terminal coding agent built on the Deep Agents SDK - works with any model, with persistent memory, customizable skills, subagents, and approval controls.
 Project-URL: Homepage, https://docs.langchain.com/oss/python/deepagents/overview
 Project-URL: Documentation, https://reference.langchain.com/python/deepagents/
 Project-URL: Repository, https://github.com/langchain-ai/deepagents
@@ -33,12 +33,12 @@ Requires-Dist: httpx<1.0.0,>=0.28.1
 Requires-Dist: langchain-anthropic<2.0.0,>=1.4.8
 Requires-Dist: langchain-google-genai<5.0.0,>=4.2.7
 Requires-Dist: langchain-mcp-adapters<1.0.0,>=0.3.0
-Requires-Dist: langchain-openai<2.0.0,>=1.3.4
+Requires-Dist: langchain-openai<2.0.0,>=1.3.5
 Requires-Dist: langchain-quickjs<0.4.0,>=0.3.2
-Requires-Dist: langchain<2.0.0,>=1.3.12
+Requires-Dist: langchain<2.0.0,>=1.3.13
 Requires-Dist: langgraph-checkpoint-sqlite<4.0.0,>=3.1.0
 Requires-Dist: langgraph-cli[inmem]<1.0.0,>=0.4.30
-Requires-Dist: langgraph-runtime-inmem<1.0.0,>=0.30.0
+Requires-Dist: langgraph-runtime-inmem<1.0.0,>=0.31.0
 Requires-Dist: langgraph-sdk<1.0.0,>=0.4.2
 Requires-Dist: langsmith[sandbox]>=0.9.3
 Requires-Dist: markdownify<2.0.0,>=1.2.2
@@ -74,10 +74,11 @@ Requires-Dist: langchain-groq<2.0.0,>=1.1.3; extra == 'all-providers'
 Requires-Dist: langchain-huggingface<2.0.0,>=1.2.2; extra == 'all-providers'
 Requires-Dist: langchain-ibm<2.0.0,>=1.1.0; extra == 'all-providers'
 Requires-Dist: langchain-litellm<2.0.0,>=0.7.0; extra == 'all-providers'
+Requires-Dist: langchain-meta<2.0.0,>=0.5.0; extra == 'all-providers'
 Requires-Dist: langchain-mistralai<2.0.0,>=1.1.6; extra == 'all-providers'
 Requires-Dist: langchain-nvidia-ai-endpoints<2.0.0,>=1.4.3; extra == 'all-providers'
 Requires-Dist: langchain-ollama<2.0.0,>=1.1.0; extra == 'all-providers'
-Requires-Dist: langchain-openai<2.0.0,>=1.3.4; extra == 'all-providers'
+Requires-Dist: langchain-openai<2.0.0,>=1.3.5; extra == 'all-providers'
 Requires-Dist: langchain-openrouter<2.0.0,>=0.2.6; extra == 'all-providers'
 Requires-Dist: langchain-perplexity<2.0.0,>=1.4.0; extra == 'all-providers'
 Requires-Dist: langchain-together<2.0.0,>=0.4.0; extra == 'all-providers'
@@ -115,6 +116,8 @@ Requires-Dist: langchain-litellm<2.0.0,>=0.7.0; extra == 'litellm'
 Provides-Extra: media
 Requires-Dist: av<18.0.0,>=17.0.0; extra == 'media'
 Requires-Dist: pillow<13.0.0,>=12.2.0; extra == 'media'
+Provides-Extra: meta
+Requires-Dist: langchain-meta<2.0.0,>=0.5.0; extra == 'meta'
 Provides-Extra: mistralai
 Requires-Dist: langchain-mistralai<2.0.0,>=1.1.6; extra == 'mistralai'
 Provides-Extra: modal
@@ -125,7 +128,7 @@ Requires-Dist: langchain-nvidia-ai-endpoints<2.0.0,>=1.4.3; extra == 'nvidia'
 Provides-Extra: ollama
 Requires-Dist: langchain-ollama<2.0.0,>=1.1.0; extra == 'ollama'
 Provides-Extra: openai
-Requires-Dist: langchain-openai<2.0.0,>=1.3.4; extra == 'openai'
+Requires-Dist: langchain-openai<2.0.0,>=1.3.5; extra == 'openai'
 Provides-Extra: openrouter
 Requires-Dist: langchain-openrouter<2.0.0,>=0.2.6; extra == 'openrouter'
 Provides-Extra: perplexity

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/__init__.py

@@ -6,12 +6,14 @@ import logging
 from typing import TYPE_CHECKING
 
 from deepagents_code._debug import configure_debug_logging
+from deepagents_code._debug_buffer import install_log_buffer
 from deepagents_code._version import __version__
 
 if TYPE_CHECKING:
     from collections.abc import Callable
 
-configure_debug_logging(logging.getLogger(__name__))  # noqa: RUF067  # package logger must be configured before child modules emit logs
+install_log_buffer(logging.getLogger(__name__))  # noqa: RUF067  # attach the always-on tail first so warnings from configure_debug_logging are captured
+configure_debug_logging(logging.getLogger(__name__))  # noqa: RUF067  # package logger must be configured before child modules emit logs; sets the final level over the buffer's INFO floor
 
 __all__ = [
     "__version__",

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/_build_info.py

@@ -1,3 +1,3 @@
 """Generated at build time. Do not edit or commit."""
 
-BUILD_COMMIT = "2f56309"
+BUILD_COMMIT = "def6369"

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/_constants.py

@@ -13,6 +13,24 @@ from typing import Final
 DEFAULT_AGENT_NAME: Final[str] = "agent"
 """Default agent / assistant identifier when no `-a` flag is given."""
 
+FIREWORKS_PROVIDER_ID_PREFIX: Final[str] = "accounts/fireworks/"
+"""Prefix used to infer Fireworks from fully-qualified IDs."""
+
+FIREWORKS_MODEL_ID_PREFIXES: Final[tuple[str, ...]] = (
+    "accounts/fireworks/models/",
+    "accounts/fireworks/routers/",
+)
+"""Model and router ID prefixes used for stripping and classification."""
+
+MCP_REENABLED_PENDING_ERROR: Final[str] = "Re-enabled — press Ctrl+R to load."
+"""User-facing reconnect guidance shown for an MCP server that was optimistically
+re-enabled but whose agent has not yet reconnected.
+
+Set as `MCPServerInfo.error` by `app._apply_optimistic_disabled_state` (alongside
+`pending_reconnect=True`, which is what `/tools` actually keys off). Named here
+so the producer and the tests asserting the message share one literal.
+"""
+
 SYSTEM_MESSAGE_PREFIX: Final[str] = "[SYSTEM]"
 """Prefix for synthetic human messages (e.g. interrupt cancellation notices).
 

deepagents_code-0.1.37/deepagents_code/_debug.py

@@ -0,0 +1,148 @@
+"""Shared debug-logging configuration for runtime and file-based tracing.
+
+When the `DEEPAGENTS_CODE_DEBUG` environment variable is set, modules that handle
+streaming or remote communication can enable detailed file-based logging. This
+helper centralizes the setup so the env-var names, file path, log level, and
+format are defined in one place.
+"""
+
+from __future__ import annotations
+
+import logging
+import os
+import sys
+from pathlib import Path
+
+from deepagents_code._env_vars import (
+    DEBUG,
+    DEBUG_FILE,
+    DEFAULT_DEBUG_FILE,
+    LOG_LEVEL,
+    is_env_truthy,
+)
+
+logger = logging.getLogger(__name__)
+
+_DEBUG_HANDLER_ATTR = "_deepagents_code_debug_handler"
+LOG_LEVELS = {
+    "DEBUG": logging.DEBUG,
+    "INFO": logging.INFO,
+    "WARNING": logging.WARNING,
+    "ERROR": logging.ERROR,
+    "CRITICAL": logging.CRITICAL,
+}
+"""Canonical level-name to `logging` level mapping.
+
+The single source of truth for level names and their numeric values, shared with
+the Debug Console's level filter so severity ordering is never re-derived from
+hardcoded integers.
+"""
+
+
+def resolve_log_level(*, debug_enabled: bool | None = None) -> int:
+    """Resolve the configured runtime logging level.
+
+    Args:
+        debug_enabled: Whether `DEEPAGENTS_CODE_DEBUG` is truthy. When omitted,
+            the current environment is checked.
+
+    Returns:
+        A standard `logging` level integer. Defaults to `DEBUG` when debug file
+        logging is enabled and `INFO` otherwise.
+    """
+    if debug_enabled is None:
+        debug_enabled = is_env_truthy(DEBUG)
+    fallback = logging.DEBUG if debug_enabled else logging.INFO
+    raw = os.environ.get(LOG_LEVEL)
+    if raw is None or not raw.strip():
+        return fallback
+    level = LOG_LEVELS.get(raw.strip().upper())
+    if level is not None:
+        return level
+    valid = ", ".join(LOG_LEVELS)
+    message = f"ignoring invalid {LOG_LEVEL}={raw!r}; expected one of {valid}"
+    # stderr for headless / pre-TUI visibility; the logger so it also lands in
+    # the always-on in-memory buffer and surfaces in the Debug Console (the
+    # buffer is installed before this runs; see __init__.py).
+    print(f"Warning: {message}", file=sys.stderr)  # noqa: T201
+    logger.warning("%s", message)
+    return fallback
+
+
+def configure_debug_logging(target: logging.Logger) -> None:
+    """Configure runtime log level and optional file logging for *target*.
+
+    Intended to be called once on the `deepagents_code` package logger; child
+    module loggers reach the same handlers via propagation, so individual modules
+    do not configure logging themselves.
+
+    `DEEPAGENTS_CODE_LOG_LEVEL` controls the package logger level independently of
+    file logging. If it is unset, `DEEPAGENTS_CODE_DEBUG=1` defaults to `DEBUG`;
+    otherwise the runtime level defaults to `INFO`.
+
+    When `DEEPAGENTS_CODE_DEBUG` is truthy, a file handler is attached. The log
+    file defaults to `DEFAULT_DEBUG_FILE` but can be overridden with
+    `DEEPAGENTS_CODE_DEBUG_FILE`. The handler appends (`mode='a'`) so logs are
+    preserved across separate process runs. Calling this again with the same
+    resolved path does not stack duplicate handlers: the existing tagged handler
+    is reused and its level re-applied. If the resolved path changes, the stale
+    handler is closed and replaced.
+
+    Args:
+        target: Logger to configure.
+    """
+    debug_enabled = is_env_truthy(DEBUG)
+    level = resolve_log_level(debug_enabled=debug_enabled)
+    target.setLevel(level)
+
+    if not debug_enabled:
+        return
+
+    debug_path = Path(os.environ.get(DEBUG_FILE, DEFAULT_DEBUG_FILE))
+    for existing in list(target.handlers):
+        if not (
+            isinstance(existing, logging.FileHandler)
+            and getattr(existing, _DEBUG_HANDLER_ATTR, False)
+        ):
+            continue
+        if Path(existing.baseFilename) == debug_path:
+            existing.setLevel(level)
+            return
+        # The debug path changed; drop the stale handler before re-attaching so
+        # we don't leak its file descriptor or fan logs out to two files.
+        target.removeHandler(existing)
+        existing.close()
+
+    try:
+        handler = logging.FileHandler(str(debug_path), mode="a")
+    except OSError as exc:
+        message = f"could not open debug log file {debug_path}: {exc}"
+        # stderr for headless / pre-TUI visibility; the logger so it also lands
+        # in the always-on in-memory buffer and surfaces in the Debug Console.
+        print(f"Warning: {message}", file=sys.stderr)  # noqa: T201
+        logger.warning("%s", message)
+        return
+    setattr(handler, _DEBUG_HANDLER_ATTR, True)
+    handler.setLevel(level)
+    handler.setFormatter(logging.Formatter("%(asctime)s %(name)s %(message)s"))
+    target.addHandler(handler)
+
+
+def installed_debug_log_path() -> Path | None:
+    """Return the path of the active debug log file, or `None` if not logging.
+
+    Reflects the file handler actually attached by `configure_debug_logging`,
+    not the current `DEEPAGENTS_CODE_DEBUG` env value. The two diverge when the
+    variable is set after import — e.g. via a project/global `.env` loaded during
+    settings bootstrap — in which case the variable reads truthy but no handler
+    was installed and no log file exists. Callers that surface "full error in
+    <path>" hints must use this rather than the env var to avoid pointing users
+    at a file that was never created.
+    """
+    package_logger = logging.getLogger(__package__ or "deepagents_code")
+    for handler in package_logger.handlers:
+        if isinstance(handler, logging.FileHandler) and getattr(
+            handler, _DEBUG_HANDLER_ATTR, False
+        ):
+            return Path(handler.baseFilename)
+    return None

deepagents_code-0.1.37/deepagents_code/_debug_buffer.py

@@ -0,0 +1,204 @@
+r"""In-memory ring buffer of recent log records for the in-app Debug Console.
+
+A lightweight `logging.Handler` keeps the most recent structured log records in a
+bounded `deque` so the Debug Console (`Ctrl+\`) can show a live tail without
+requiring the opt-in file logging from `_debug.configure_debug_logging`. The
+handler is installed once on the `deepagents_code` package logger (see
+`__init__.py`); child module loggers reach it via propagation.
+
+Installation is negligible, and each emitted record only appends a structured
+record to a bounded `deque`, so the buffer is cheap enough to keep always on.
+"""
+
+from __future__ import annotations
+
+import itertools
+import logging
+import os
+from collections import deque
+from dataclasses import dataclass
+
+from deepagents_code._env_vars import LOG_LEVEL
+
+DEFAULT_CAPACITY = 1000
+"""Maximum number of records retained in the ring buffer."""
+
+_DATE_FORMAT = "%H:%M:%S"
+_FORMATTER = logging.Formatter(datefmt=_DATE_FORMAT)
+
+
+@dataclass(frozen=True, slots=True)
+class InMemoryLogRecord:
+    """Structured log record retained by the in-memory debug buffer."""
+
+    timestamp: str
+    level: str
+    levelno: int
+    logger: str
+    message: str
+
+    @property
+    def plain_line(self) -> str:
+        """The record in the legacy plain-text debug-console format."""
+        return f"{self.timestamp} {self.level} {self.logger} {self.message}"
+
+
+class InMemoryLogBuffer(logging.Handler):
+    """Logging handler retaining the most recent structured records in memory."""
+
+    def __init__(self, capacity: int = DEFAULT_CAPACITY) -> None:
+        """Create the handler with a bounded backing `deque`.
+
+        Args:
+            capacity: Maximum records to retain; oldest are dropped first.
+
+        Raises:
+            ValueError: If *capacity* is less than 1. A zero-capacity buffer
+                would silently discard every record while `total_emitted` kept
+                climbing, so it is rejected at the boundary.
+        """
+        if capacity < 1:
+            msg = f"capacity must be >= 1, got {capacity}"
+            raise ValueError(msg)
+        super().__init__()
+        self._records: deque[InMemoryLogRecord] = deque(maxlen=capacity)
+        self._total = 0
+
+    def emit(self, record: logging.LogRecord) -> None:
+        """Append the structured *record* to the ring buffer."""
+        self.acquire()
+        try:
+            self._records.append(self._make_record(record))
+            self._total += 1
+        except Exception:  # noqa: BLE001  # never let logging crash the app
+            self.handleError(record)
+        finally:
+            self.release()
+
+    @property
+    def total_emitted(self) -> int:
+        """Total records ever emitted (monotonic; survives buffer eviction)."""
+        self.acquire()
+        try:
+            return self._total
+        finally:
+            self.release()
+
+    def snapshot_records_since(self, index: int) -> tuple[list[InMemoryLogRecord], int]:
+        """Return retained structured records and the next absolute index.
+
+        Absolute indices are stable even as old records are evicted, so callers
+        can poll incrementally without re-reading records they already consumed.
+        Returning the records and the resume index together under the handler
+        lock prevents a concurrent append from being skipped between separate
+        reads.
+
+        Args:
+            index: Absolute emission index to start from.
+
+        Returns:
+            A tuple of the structured records from *index* onward that are still
+            retained, and the current total emitted count to resume from.
+        """
+        self.acquire()
+        try:
+            return self._records_since_unlocked(index), self._total
+        finally:
+            self.release()
+
+    def _records_since_unlocked(self, index: int) -> list[InMemoryLogRecord]:
+        """Return retained records from *index* while the handler lock is held."""
+        start_abs = self._total - len(self._records)
+        offset = index - start_abs
+        if offset <= 0:
+            return list(self._records)
+        if offset >= len(self._records):
+            return []
+        # islice avoids materializing the full deque just to slice off the tail.
+        return list(itertools.islice(self._records, offset, None))
+
+    @staticmethod
+    def _make_record(record: logging.LogRecord) -> InMemoryLogRecord:
+        """Convert a standard logging record into a structured debug record.
+
+        Returns:
+            Structured record for display and filtering.
+        """
+        message = record.getMessage()
+        if record.exc_info:
+            if not record.exc_text:
+                record.exc_text = _FORMATTER.formatException(record.exc_info)
+            if record.exc_text:
+                message = f"{message}\n{record.exc_text}"
+        if record.stack_info:
+            message = f"{message}\n{_FORMATTER.formatStack(record.stack_info)}"
+        return InMemoryLogRecord(
+            timestamp=_FORMATTER.formatTime(record, _DATE_FORMAT),
+            level=record.levelname,
+            levelno=record.levelno,
+            logger=record.name,
+            message=message,
+        )
+
+
+_buffer: InMemoryLogBuffer | None = None
+
+
+def install_log_buffer(
+    target: logging.Logger, capacity: int = DEFAULT_CAPACITY
+) -> InMemoryLogBuffer:
+    """Attach the in-memory buffer handler to *target* (idempotent).
+
+    Lowers *target*'s level to at most `INFO` so the console shows a useful tail
+    even when `DEEPAGENTS_CODE_DEBUG` is off; never raises the level. In
+    `__init__.py` this runs *before* `configure_debug_logging`, which then sets
+    the final level (honoring `DEEPAGENTS_CODE_DEBUG` and
+    `DEEPAGENTS_CODE_LOG_LEVEL`) over this `INFO` floor — so any startup warnings
+    `configure_debug_logging` emits are captured by the already-installed buffer.
+    On a fresh `NOTSET` logger the `NOTSET` branch forces `INFO` without
+    consulting `DEEPAGENTS_CODE_LOG_LEVEL`; the `> INFO and no env` branch only
+    matters on reconfiguration (e.g. an `importlib.reload`), where it preserves
+    an explicit `DEEPAGENTS_CODE_LOG_LEVEL` rather than clobbering it with `INFO`.
+
+    Lowering the level does not spill log output onto the terminal: because this
+    handler is present in the propagation chain, `Logger.callHandlers` finds a
+    handler (`found > 0`) and Python's `lastResort` stderr handler is never
+    consulted. The exception is an embedding process that attaches its own
+    `INFO`-or-lower handler to the root logger, which would then see the
+    propagated records.
+
+    Note: this runs as an import-time side effect (see `__init__.py`), so every
+    `import deepagents_code` attaches the handler and may lower the package
+    logger's level to `INFO` for the lifetime of the process.
+
+    Args:
+        target: Logger to attach the buffer to (the package logger).
+        capacity: Maximum records to retain.
+
+    Returns:
+        The installed (or already-installed) buffer handler.
+    """
+    global _buffer  # noqa: PLW0603  # module-level singleton accessor
+    for existing in target.handlers:
+        if isinstance(existing, InMemoryLogBuffer):
+            _buffer = existing
+            return existing
+
+    handler = InMemoryLogBuffer(capacity)
+    handler.setLevel(logging.DEBUG)
+    target.addHandler(handler)
+    if target.level == logging.NOTSET or (
+        target.level > logging.INFO and not os.environ.get(LOG_LEVEL)
+    ):
+        target.setLevel(logging.INFO)
+    _buffer = handler
+    return handler
+
+
+def get_log_buffer() -> InMemoryLogBuffer | None:
+    """Return the installed buffer handler.
+
+    Returns:
+        The installed buffer handler, or `None` if not yet installed.
+    """
+    return _buffer

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/_env_vars.py

@@ -44,6 +44,13 @@ rather than forcing the default). Also settable via `[ui].collapse_pastes` in
 config.toml.
 """
 
+CURSOR_STYLE = "DEEPAGENTS_CODE_CURSOR_STYLE"
+"""Chat input cursor style (`block` or `underline`).
+
+Takes precedence over `[ui].cursor_style` in config.toml. Invalid values fall
+through to the config file and then the default block cursor.
+"""
+
 DEBUG = "DEEPAGENTS_CODE_DEBUG"
 """Enable verbose debug logging and preserve the server subprocess log.
 
@@ -183,6 +190,12 @@ The value is comma-separated for forward-compatibility, not because multiple
 destinations are written today.
 """
 
+LOG_LEVEL = "DEEPAGENTS_CODE_LOG_LEVEL"
+"""Minimum level for `deepagents_code` runtime logging.
+
+Accepted values are DEBUG, INFO, WARNING, ERROR, and CRITICAL.
+"""
+
 NO_TERMINAL_ESCAPE = "DEEPAGENTS_CODE_NO_TERMINAL_ESCAPE"
 """Disable all terminal escape/control sequence output when enabled."""
 

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/_version.py

@@ -2,7 +2,7 @@
 
 # Keep the `x-release-please-version` annotation — release-please uses it to
 # bump `__version__` in sync with `pyproject.toml` on every release PR.
-__version__ = "0.1.36"  # x-release-please-version
+__version__ = "0.1.37"  # x-release-please-version
 
 DOCS_URL = "https://docs.langchain.com/oss/python/deepagents/code"
 """URL for `deepagents-code` documentation."""

{deepagents_code-0.1.36 → deepagents_code-0.1.37}/deepagents_code/agent.py

@@ -1347,8 +1347,18 @@ def create_cli_agent(
             Used for system prompt generation.
         system_prompt: Override the default system prompt.
 
-            If `None`, generates one based on `sandbox_type`, `assistant_id`,
-            and `interactive`.
+            If `None`, a system prompt is auto-generated with dynamic context
+            interpolated in (model identity, working directory, sandbox vs.
+            local execution mode, skills path, and interactive-vs-headless
+            guidance).
+
+            !!! warning
+
+                Passing a value here replaces that auto-generated prompt
+                entirely — none of the dynamic context above is added, and
+                `sandbox_type` and `interactive` no longer influence the
+                prompt. Only pass an explicit prompt when you intend to take
+                full ownership of the system prompt's content.
         interactive: When `False`, the auto-generated system prompt is
             tailored for headless non-interactive execution. Ignored when
             `system_prompt` is provided explicitly.