codeer-cli 0.1.8__tar.gz → 0.1.10__tar.gz

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/API_REFERENCE.md

@@ -49,6 +49,23 @@ like `knowledge_node_ids`, `domain`, `agent_id`, `http_request` config.
 
 Limits: 10 tools per agent, ≤5 `call_agent`, ≤1 `memory`.
 
+Agent create/update payloads may also include:
+
+```json
+{
+  "human_handoff": {
+    "enabled": true,
+    "idle_timeout_minutes": null,
+    "handoff_instructions": "Hand off when the user asks for a person."
+  }
+}
+```
+
+`idle_timeout_minutes` must be greater than zero when provided. The runtime
+only exposes the handoff tool to an external conversation with a non-empty
+`external_user_id`, or to an evaluation run. Internal editor Live Test does not
+activate human mode.
+
 ## Stage 2 — Knowledge bases
 
 Base path: `/organizations/{org_id}/workspaces/{ws_id}/knowledge_bases`
@@ -149,10 +166,10 @@ for the apply → test → publish workflow. Pass the draft `AgentHistory.id` fr
 
 | Method & path | Purpose |
 | --- | --- |
-| `GET /eval/workspaces/{workspace_id}/case-labels` | List reusable eval case labels |
-| `POST /eval/workspaces/{workspace_id}/case-labels` | Create reusable eval case label (`name`, `color?`) |
-| `PUT /eval/case-labels/{label_id}` | Update eval case label name/color |
-| `DELETE /eval/case-labels/{label_id}` | Delete eval case label and clear associations |
+| `GET /external/eval/case-labels` | List reusable eval case labels in the API key workspace |
+| `POST /external/eval/case-labels` | Create reusable eval case label (`name`, `color?`) in the API key workspace |
+| `PUT /external/eval/case-labels/{label_id}` | Update an eval case label in the API key workspace |
+| `DELETE /external/eval/case-labels/{label_id}` | Delete an eval case label and clear associations |
 | `POST /eval/cases` | Create case (`input`, `expected_output?`, `rubric?`, `label_ids?`); rubric = user-docs "Standard" |
 | `GET /eval/agents/{agent_id}/cases` | List cases for an agent |
 | `GET /eval/cases/{case_id}` | Read one |
@@ -171,8 +188,11 @@ for the apply → test → publish workflow. Pass the draft `AgentHistory.id` fr
 | `POST /eval/rubric` | Set/override the rubric for one (case, evaluator); also creates assignment |
 | `POST /eval/rubrics/batch` | **Read** rubrics for a batch of (case, evaluator) pairs |
 
-Eval case labels are workspace-scoped reusable objects. The case create/update
-payload uses `label_ids` (stringified label IDs), not freeform label names:
+Eval case labels are workspace-scoped reusable objects. The external label
+endpoints infer the workspace from the API key and do not accept a workspace
+override; switch CLI profiles to operate on another workspace. The case
+create/update payload uses `label_ids` (stringified label IDs), not freeform
+label names:
 
 ```json
 {
@@ -247,12 +267,19 @@ the public CLI.
 
 | Method & path | Purpose |
 | --- | --- |
+| `POST /chats` | Create a persisted history using an agent's current published version |
+| `POST /chats/{id}/messages` | Append a turn to an existing persisted history using the current published version |
 | `GET /histories?agent_id=X&feedback_filter=improve_feedback&external_user_id=…` | List conversations with filters |
 | `GET /histories/{id}` | Read one history's metadata |
 | `GET /histories/{id}/conversations` | Full conversation turns incl. tool calls |
 | `POST /histories/{hid}/conversations/{cid}/feedbacks` | Leave freeform improvement feedback |
 | `POST /histories/{hid}/conversations/{cid}/score` | Numeric score |
 
+The CLI exposes the first two operations as `codeer history create` and
+`codeer history send`. Non-streaming message requests default to a 120-second
+per-message timeout. A timeout has an uncertain write outcome, so read the
+history before retrying to avoid duplicate turns.
+
 `feedback_filter` accepts the `FeedbackFilterType` enum values:
 `no_feedback`, `with_feedback`, `helpful_feedback`, `improve_feedback`.
 
@@ -270,7 +297,7 @@ Non-destructive: older versions stay in `GET /agents/{id}/histories`.
 | --- | --- |
 | `GET /accounts/me` | Sanity-check session, read workspace_organization_map |
 | `GET /organizations` | List orgs visible to the user |
-| `GET /llm/models` | List available LLM model IDs to use as `llm_model` |
+| `GET /llm/models` | List available LLM model IDs to use as `llm_model` (`codeer model list`) |
 | `GET /retrieval/...` | Shared retrieval helpers (file upload for attachments, markdown conversion) |
 
 ---

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codeer-cli
-Version: 0.1.8
+Version: 0.1.10
 Summary: Command line tools for managing Codeer agents over the Codeer API.
 Project-URL: Homepage, https://www.codeer.ai
 Author: Codeer.AI
@@ -117,6 +117,35 @@ Validate setup before API work:
 codeer check
 ```
 
+List the active cloud models without opening the Codeer web app:
+
+```bash
+codeer model list --type text
+```
+
+## Agent human handoff
+
+`codeer agent apply` accepts the same `human_handoff` object as the Agent API.
+The dry-run validates it and shows whether handoff is enabled before any server
+write:
+
+```json
+{
+  "name": "Support Agent",
+  "system_prompt": "Help the user safely.",
+  "human_handoff": {
+    "enabled": true,
+    "idle_timeout_minutes": null,
+    "handoff_instructions": "Hand off when the user asks for a person."
+  }
+}
+```
+
+`idle_timeout_minutes` must be a positive integer or `null`. Human handoff only
+becomes available in live published-agent conversations with a non-empty
+`external_user_id`; editor Live Test conversations are internal and cannot
+activate human mode.
+
 ## Upgrade and uninstall
 
 Upgrade the CLI:
@@ -143,9 +172,20 @@ Use this pattern during agent lifecycle work:
 ```bash
 codeer agent list
 codeer history list --agent <agent-id> --limit 50
+codeer history create --agent <agent-id> --message "Review this plan" --timeout 120
+codeer history send <history-id> --message "Use the recommended options" --timeout 120
 codeer eval run --agent <agent-id> --cases <case-ids> --evaluator <evaluator-id> --out .codeer/eval_run.json
 ```
 
+`history create` and `history send` use the agent's current published version.
+Their per-message timeout defaults to 120 seconds. If a write request times
+out, inspect the history before retrying: the server may have completed the
+turn after the client stopped waiting.
+
+Eval case label commands always operate on the active API-key workspace. They
+do not accept a workspace override; switch CLI profiles to target another
+workspace.
+
 Flags:
 
 - `--full` prints bounded extra detail for human inspection. It is still

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/README.md

@@ -99,6 +99,35 @@ Validate setup before API work:
 codeer check
 ```
 
+List the active cloud models without opening the Codeer web app:
+
+```bash
+codeer model list --type text
+```
+
+## Agent human handoff
+
+`codeer agent apply` accepts the same `human_handoff` object as the Agent API.
+The dry-run validates it and shows whether handoff is enabled before any server
+write:
+
+```json
+{
+  "name": "Support Agent",
+  "system_prompt": "Help the user safely.",
+  "human_handoff": {
+    "enabled": true,
+    "idle_timeout_minutes": null,
+    "handoff_instructions": "Hand off when the user asks for a person."
+  }
+}
+```
+
+`idle_timeout_minutes` must be a positive integer or `null`. Human handoff only
+becomes available in live published-agent conversations with a non-empty
+`external_user_id`; editor Live Test conversations are internal and cannot
+activate human mode.
+
 ## Upgrade and uninstall
 
 Upgrade the CLI:
@@ -125,9 +154,20 @@ Use this pattern during agent lifecycle work:
 ```bash
 codeer agent list
 codeer history list --agent <agent-id> --limit 50
+codeer history create --agent <agent-id> --message "Review this plan" --timeout 120
+codeer history send <history-id> --message "Use the recommended options" --timeout 120
 codeer eval run --agent <agent-id> --cases <case-ids> --evaluator <evaluator-id> --out .codeer/eval_run.json
 ```
 
+`history create` and `history send` use the agent's current published version.
+Their per-message timeout defaults to 120 seconds. If a write request times
+out, inspect the history before retrying: the server may have completed the
+turn after the client stopped waiting.
+
+Eval case label commands always operate on the active API-key workspace. They
+do not accept a workspace override; switch CLI profiles to target another
+workspace.
+
 Flags:
 
 - `--full` prints bounded extra detail for human inspection. It is still

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "codeer-cli"
-version = "0.1.8"
+version = "0.1.10"
 description = "Command line tools for managing Codeer agents over the Codeer API."
 readme = "README.md"
 requires-python = ">=3.11"

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/__init__.py

@@ -18,8 +18,8 @@ production and can be overridden from process env. The CLI does not read
 workspace-local dotenv files or credential files.
 """
 
-from ._validate import ToolValidationError
-from .client import AuthError, CodeerClient, CodeerError
+from ._validate import HumanHandoffValidationError, ToolValidationError
+from .client import AuthError, CodeerClient, CodeerError, TransportError
 from .parse import (
     AgentSummary,
     ConversationTurn,
@@ -43,7 +43,8 @@ from .parse import (
 )
 
 __all__ = [
-    "CodeerClient", "CodeerError", "AuthError", "ToolValidationError",
+    "CodeerClient", "CodeerError", "TransportError", "AuthError", "ToolValidationError",
+    "HumanHandoffValidationError",
     # parsers
     "AgentSummary", "ConversationTurn", "EvalResultSummary", "HistorySummary",
     "KBNode", "ToolCall", "EvalToolCall",

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/_validate.py

@@ -1,4 +1,4 @@
-"""Client-side validation for unified-tool payloads.
+"""Client-side validation for agent payloads.
 
 These checks exist because the backend's form-schema validator is lenient
 (``extra="allow"``) and silently accepts unknown ``type`` strings, which then
@@ -24,6 +24,47 @@ class ToolValidationError(ValueError):
     """Raised when a unified_tools payload is definitely wrong."""
 
 
+class HumanHandoffValidationError(ValueError):
+    """Raised when a human_handoff payload is definitely wrong."""
+
+
+def validate_human_handoff(config: Any) -> dict[str, Any] | None:
+    """Validate and normalize an optional human-handoff configuration."""
+    if config is None:
+        return None
+    if not isinstance(config, dict):
+        raise HumanHandoffValidationError("human_handoff must be an object.")
+
+    allowed_keys = {"enabled", "idle_timeout_minutes", "handoff_instructions"}
+    unknown_keys = sorted(set(config) - allowed_keys)
+    if unknown_keys:
+        raise HumanHandoffValidationError(
+            f"human_handoff contains unsupported field(s): {', '.join(unknown_keys)}."
+        )
+
+    enabled = config.get("enabled", False)
+    if not isinstance(enabled, bool):
+        raise HumanHandoffValidationError("human_handoff.enabled must be true or false.")
+
+    timeout = config.get("idle_timeout_minutes")
+    if timeout is not None and (type(timeout) is not int or timeout <= 0):
+        raise HumanHandoffValidationError(
+            "human_handoff.idle_timeout_minutes must be a positive integer or null."
+        )
+
+    instructions = config.get("handoff_instructions")
+    if instructions is not None and not isinstance(instructions, str):
+        raise HumanHandoffValidationError(
+            "human_handoff.handoff_instructions must be a string or null."
+        )
+
+    return {
+        "enabled": enabled,
+        "idle_timeout_minutes": timeout,
+        "handoff_instructions": instructions,
+    }
+
+
 def validate_unified_tools(tools: Iterable[dict[str, Any]] | None) -> list[dict[str, Any]]:
     """Raise ToolValidationError on invalid payloads; otherwise return the list.
 

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/agents.py

@@ -7,7 +7,7 @@ from __future__ import annotations
 
 from typing import Any, List, Optional
 
-from ._validate import validate_unified_tools
+from ._validate import validate_human_handoff, validate_unified_tools
 from .client import CodeerClient
 
 
@@ -24,8 +24,10 @@ def create(
     suggested_questions: Optional[List[str]] = None,
     primary_object_ids: Optional[List[int]] = None,
     attachment_ids: Optional[List[str]] = None,
+    human_handoff: Optional[dict[str, Any]] = None,
 ) -> dict:
     validated_tools = validate_unified_tools(unified_tools)
+    validated_handoff = validate_human_handoff(human_handoff)
     body: dict[str, Any] = {
         "name": name,
         "system_prompt": system_prompt,
@@ -39,6 +41,8 @@ def create(
         body["description"] = description
     if llm_model is not None:
         body["llm_model"] = llm_model
+    if validated_handoff is not None:
+        body["human_handoff"] = validated_handoff
     return client.post("/external/agents", json=body)
 
 
@@ -56,9 +60,11 @@ def update(
     suggested_questions: Optional[List[str]] = None,
     primary_object_ids: Optional[List[int]] = None,
     attachment_ids: Optional[List[str]] = None,
+    human_handoff: Optional[dict[str, Any]] = None,
 ) -> dict:
     """PUT creates a new AgentHistory snapshot (draft)."""
     validated_tools = validate_unified_tools(unified_tools)
+    validated_handoff = validate_human_handoff(human_handoff)
     body: dict[str, Any] = {
         "name": name,
         "system_prompt": system_prompt,
@@ -73,6 +79,8 @@ def update(
         body["description"] = description
     if llm_model is not None:
         body["llm_model"] = llm_model
+    if validated_handoff is not None:
+        body["human_handoff"] = validated_handoff
     return client.patch(f"/external/agents/{agent_id}", json=body)
 
 

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/chats.py

@@ -36,6 +36,7 @@ def send_published_agent_message(
     external_user_id: Optional[str] = None,
     attachment_ids: Optional[List[str]] = None,
     stream: bool = False,
+    timeout: Optional[float] = None,
 ) -> Iterator[dict] | dict:
     """Send a user message through the API-key external chat flow.
 
@@ -51,7 +52,7 @@ def send_published_agent_message(
     path = f"/chats/{chat_id}/messages"
     if stream:
         return client.stream_sse("POST", path, json=body)
-    return client.post(path, json=body)
+    return client.post(path, json=body, timeout=timeout)
 
 
 def send_message(
@@ -84,4 +85,3 @@ def list_messages(client: CodeerClient, chat_id: int) -> list[dict]:
 
 def list_chats(client: CodeerClient) -> list[dict]:
     return client.get("/chats")
-

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/cli.py

@@ -2,9 +2,10 @@
 
     codeer check
     codeer agent list|get|apply|diff|versions
+    codeer model list
     codeer kb list|files|upload|node-rename|node-delete|faq-list|faq-get|faq-create|faq-update|faq-delete
     codeer eval list|label-list|label-create|label-update|label-delete|case-update|case-delete|evaluators|evaluator-create|evaluator-update|run|export|reconcile|cases-apply|rubrics|rubrics-apply
-    codeer history list|get|conversations|negative-feedback
+    codeer history list|get|conversations|negative-feedback|create|send
 """
 
 from __future__ import annotations
@@ -25,6 +26,7 @@ def main(argv: list[str] | None = None) -> int:
         epilog="""\
 Safe workflow for coding agents:
   codeer check --json
+  codeer model list --type text
   codeer agent list
   codeer agent get <agent-id> --full
   codeer kb list
@@ -55,6 +57,12 @@ Use --out <path> for large raw artifacts; stdout defaults to compact summaries.
 
     check.register(sub)
 
+    try:
+        from .commands import model as model_cmd
+        model_cmd.register(sub)
+    except ImportError:
+        pass
+
     # Phase 2-4: agent, kb, eval commands will register here
     try:
         from .commands import agent as agent_cmd

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/client.py

@@ -37,6 +37,16 @@ class ScopeResolutionError(CodeerError):
     """Raised when workspace or organization scope cannot be inferred."""
 
 
+class TransportError(CodeerError):
+    """Raised when an HTTP request fails before a response is available."""
+
+    def __init__(self, message: str, body: Any = None):
+        RuntimeError.__init__(self, message)
+        self.status = 0
+        self.message = message
+        self.body = body
+
+
 @dataclass
 class CodeerClient:
     """Thin wrapper around httpx.Client with Codeer API-key auth.
@@ -148,9 +158,32 @@ class CodeerClient:
         json: Any = None,
         files: Any = None,
         data: Any = None,
+        timeout: Optional[float] = None,
     ) -> Any:
         url = path if path.startswith("http") else f"/api/v1{path if path.startswith('/') else '/' + path}"
-        r = self._client.request(method, url, params=params, json=json, files=files, data=data)
+        request_kwargs: dict[str, Any] = {}
+        if timeout is not None:
+            request_kwargs["timeout"] = timeout
+        method_upper = method.upper()
+        try:
+            r = self._client.request(
+                method_upper,
+                url,
+                params=params,
+                json=json,
+                files=files,
+                data=data,
+                **request_kwargs,
+            )
+        except httpx.TimeoutException as exc:
+            raise self._transport_error(
+                method_upper,
+                path,
+                exc,
+                timeout_seconds=timeout if timeout is not None else self.timeout,
+            ) from exc
+        except httpx.RequestError as exc:
+            raise self._transport_error(method_upper, path, exc) from exc
         return self._parse(r)
 
     def get(self, path: str, **kwargs: Any) -> Any:
@@ -181,29 +214,68 @@ class CodeerClient:
         Each event is a dict like ``{"event": "message", "data": <parsed-json-or-str>}``.
         """
         url = path if path.startswith("http") else f"/api/v1{path if path.startswith('/') else '/' + path}"
-        with self._client.stream(method, url, params=params, json=json) as r:
-            if r.status_code >= 400:
-                body = r.read().decode("utf-8", "replace")
-                self._raise_for_error(r.status_code, body)
-            event = "message"
-            buf: list[str] = []
-            for line in r.iter_lines():
-                if line == "":
-                    if buf:
-                        raw = "\n".join(buf)
-                        yield {"event": event, "data": _maybe_json(raw)}
-                        buf = []
-                        event = "message"
-                    continue
-                if line.startswith(":"):
-                    continue
-                if line.startswith("event:"):
-                    event = line[len("event:"):].strip()
-                    continue
-                if line.startswith("data:"):
-                    buf.append(line[len("data:"):].lstrip())
-            if buf:
-                yield {"event": event, "data": _maybe_json("\n".join(buf))}
+        method_upper = method.upper()
+        try:
+            with self._client.stream(method_upper, url, params=params, json=json) as r:
+                if r.status_code >= 400:
+                    body = r.read().decode("utf-8", "replace")
+                    self._raise_for_error(r.status_code, body)
+                event = "message"
+                buf: list[str] = []
+                for line in r.iter_lines():
+                    if line == "":
+                        if buf:
+                            raw = "\n".join(buf)
+                            yield {"event": event, "data": _maybe_json(raw)}
+                            buf = []
+                            event = "message"
+                        continue
+                    if line.startswith(":"):
+                        continue
+                    if line.startswith("event:"):
+                        event = line[len("event:"):].strip()
+                        continue
+                    if line.startswith("data:"):
+                        buf.append(line[len("data:"):].lstrip())
+                if buf:
+                    yield {"event": event, "data": _maybe_json("\n".join(buf))}
+        except httpx.TimeoutException as exc:
+            raise self._transport_error(
+                method_upper,
+                path,
+                exc,
+                timeout_seconds=self.timeout,
+            ) from exc
+        except httpx.RequestError as exc:
+            raise self._transport_error(method_upper, path, exc) from exc
+
+    def _transport_error(
+        self,
+        method: str,
+        path: str,
+        exc: httpx.RequestError,
+        *,
+        timeout_seconds: float | None = None,
+    ) -> TransportError:
+        outcome_uncertain = (
+            method not in {"GET", "HEAD", "OPTIONS"}
+            and not isinstance(exc, httpx.ConnectError)
+        )
+        if isinstance(exc, httpx.TimeoutException):
+            timeout_value = timeout_seconds if timeout_seconds is not None else self.timeout
+            message = f"Request timed out after {timeout_value:g}s: {method} {path}."
+            if outcome_uncertain:
+                message += " The server may have completed the request; inspect current state before retrying."
+        else:
+            message = f"Request failed: {method} {path}: {exc}"
+        return TransportError(
+            message,
+            {
+                "method": method,
+                "path": path,
+                "outcome_uncertain": outcome_uncertain,
+            },
+        )
 
     def _parse(self, r: httpx.Response) -> Any:
         text = r.text

{codeer_cli-0.1.8 → codeer_cli-0.1.10}/src/codeer_cli/commands/agent.py

@@ -6,6 +6,7 @@ from pathlib import Path
 from typing import Optional
 
 from .. import agents as agents_mod
+from .._validate import validate_human_handoff, validate_unified_tools
 from ..client import CodeerClient
 from ._util import log, print_json, strip_noisy_fields, truncate, write_json
 
@@ -102,6 +103,12 @@ def _tool_summary(tools: list[dict] | None) -> list[dict]:
 
 def _agent_summary(agent: dict, *, full: bool = False) -> dict:
     tools = agent.get("unified_tools") or agent.get("tools") or []
+    human_handoff = agent.get("human_handoff")
+    if not isinstance(human_handoff, dict):
+        meta = agent.get("meta")
+        human_handoff = meta.get("human_handoff") if isinstance(meta, dict) else {}
+    if not isinstance(human_handoff, dict):
+        human_handoff = {}
     row = {
         "id": agent.get("id"),
         "name": agent.get("name"),
@@ -118,6 +125,7 @@ def _agent_summary(agent: dict, *, full: bool = False) -> dict:
         "agent_type": agent.get("agent_type"),
         "updated_at": agent.get("updated_at"),
         "tool_count": len(tools),
+        "human_handoff_enabled": bool(human_handoff.get("enabled")),
         "system_prompt_chars": len(agent.get("system_prompt") or ""),
     }
     if full:
@@ -125,6 +133,7 @@ def _agent_summary(agent: dict, *, full: bool = False) -> dict:
         row["use_search"] = agent.get("use_search")
         row["suggested_questions"] = agent.get("suggested_questions") or []
         row["tools"] = _tool_summary(tools)
+        row["human_handoff"] = human_handoff
         row["system_prompt_preview"] = truncate(agent.get("system_prompt") or "", 1200)
     return row
 
@@ -153,6 +162,13 @@ def run_apply(args, client) -> int:
         log(f"error: payload missing required field(s): {', '.join(missing)}")
         return 2
 
+    try:
+        validated_tools = validate_unified_tools(body.get("unified_tools") or [])
+        validated_handoff = validate_human_handoff(body.get("human_handoff"))
+    except ValueError as exc:
+        log(f"error: invalid agent payload: {exc}")
+        return 2
+
     if args.dry_run:
         operation = "update" if args.agent_id else "create"
         result = {
@@ -162,9 +178,14 @@ def run_apply(args, client) -> int:
             "payload": str(Path(args.payload)),
             "name": body.get("name"),
             "system_prompt_chars": len(body.get("system_prompt") or ""),
-            "tool_count": len(body.get("unified_tools") or []),
+            "tool_count": len(validated_tools),
             "use_search": body.get("use_search", False),
             "llm_model": body.get("llm_model"),
+            "human_handoff": {
+                "enabled": bool((validated_handoff or {}).get("enabled")),
+                "idle_timeout_minutes": (validated_handoff or {}).get("idle_timeout_minutes"),
+                "instructions_chars": len((validated_handoff or {}).get("handoff_instructions") or ""),
+            },
             "version_note": args.note if args.agent_id else None,
             "would_write_server_state": True,
             "next_step": "Review this summary, then rerun without --dry-run after approval.",
@@ -178,7 +199,7 @@ def run_apply(args, client) -> int:
             client, args.agent_id,
             name=body["name"],
             system_prompt=body["system_prompt"],
-            unified_tools=body.get("unified_tools") or [],
+            unified_tools=validated_tools,
             use_search=body.get("use_search", False),
             version_note=args.note,
             description=body.get("description"),
@@ -186,6 +207,7 @@ def run_apply(args, client) -> int:
             suggested_questions=body.get("suggested_questions") or [],
             primary_object_ids=body.get("primary_object_ids") or [],
             attachment_ids=body.get("attachment_ids") or [],
+            human_handoff=validated_handoff,
         )
         agent_id = args.agent_id
         log(f"PUT /agents/{agent_id} ok")
@@ -197,13 +219,14 @@ def run_apply(args, client) -> int:
             workspace_id=body["workspace_id"],
             name=body["name"],
             system_prompt=body["system_prompt"],
-            unified_tools=body.get("unified_tools") or [],
+            unified_tools=validated_tools,
             use_search=body.get("use_search", False),
             description=body.get("description"),
             llm_model=body.get("llm_model"),
             suggested_questions=body.get("suggested_questions") or [],
             primary_object_ids=body.get("primary_object_ids") or [],
             attachment_ids=body.get("attachment_ids") or [],
+            human_handoff=validated_handoff,
         )
         agent_id = agent["id"]
         log(f"POST /agents ok, id={agent_id}")