Global assistant: send a full workflow YAML only for multistep changes

.changeset/early-carrots-look.md

@@ -0,0 +1,5 @@
+---
+"apollo": major
+---
+
+enable global assistant multi step changes

pyproject.toml

@@ -10,29 +10,29 @@ requires-poetry = ">=2.3.2"
 
 [tool.poetry.dependencies]
 python = "3.11.*"
-openai = "^2.41"
+openai = "^2.44"
 python-dotenv = "^1.2.2"
-anthropic = "^0.109.1"
+anthropic = "^0.112.0"
 
 langchain-pinecone = "^0.2.13"
 langchain-core = "^1.4"
 langchain-community = "^0.4.2"
 langchain-openai = "^1.3"
 langchain-text-splitters = "^1.1"
 nltk = "^3.9.3"
-pytest = "^9.0.3"
-sentry-sdk = "^2.62.0"
+pytest = "^9.1.1"
+sentry-sdk = "^2.63.0"
 psycopg2-binary = "^2.9.10"
-langfuse = "^4.0.1"
+langfuse = "^4.12.0"
 opentelemetry-instrumentation-anthropic = "^0.61.0"
-opentelemetry-instrumentation-threading = "0.63b1"
+opentelemetry-instrumentation-threading = "0.64b0"
 
 [tool.poetry.group.dev]
 optional = false
 
 [tool.poetry.group.dev.dependencies]
-pytest = "^9.0.3"
-ruff = "^0.15.17"
+pytest = "^9.1.1"
+ruff = "^0.15.20"
 
 [build-system]
 requires = ["poetry-core"]

services/global_chat/PAYLOAD_SPEC.md

@@ -127,7 +127,7 @@ This document defines the input and output payload structure for the Global Agen
 
 - **`response`** (string): The main text response from the agent.
 
-- **`attachments`** (array): Artifacts produced during this turn. Each entry has a `type` and `content` field. An empty list `[]` means no artifacts were produced (e.g. a purely informational response). Currently supported types: `workflow_yaml`, `job_code`. When both are present, `job_code` contains the suggested code for a specific job and `workflow_yaml` contains the full YAML with the code stitched in.
+- **`attachments`** (array): Artifacts produced during this turn. Each entry has a `type` and `content` field. An empty list `[]` means no artifacts were produced (e.g. a purely informational response). The only supported type is `workflow_yaml`: the full workflow YAML with any job code changes stitched in. Job code edits are never returned separately — the YAML is the single source of truth, which allows multi-step changes in one response.
 
 - **`history`** (array): Updated conversation history including the latest exchange. On direct routes (workflow_agent, job_code_agent), each entry has `content` as a string. On the planner path, entries may have `content` as an array of content blocks (`text`, `tool_use`, `tool_result`) — this is the raw Anthropic messages format from the tool-calling loop.
 

services/global_chat/README.md

@@ -109,7 +109,9 @@ Request to build a new multi-step workflow from scratch:
 - `response`: The assistant's text response to the user
 - `attachments`: Artifacts produced — currently always
   `{type: "workflow_yaml", content: string}` when YAML was generated or modified
-- `history`: Updated conversation history including the latest exchange
+- `history`: Updated conversation history including the latest exchange. Direct
+  routes return string `content`; the planner path may return `content` as
+  Anthropic content-block arrays (`tool_use`/`tool_result`)
 - `usage`: Aggregated token usage across all agents called during the request
 - `meta.agents`: Ordered list of agents invoked (e.g.
   `["router", "workflow_agent"]` or
@@ -163,18 +165,18 @@ then calls `call_job_code_agent` for each job that needs code. Job code is
 stitched into the workflow YAML immediately after each call.
 
 The loop continues until the model signals it is done (up to a configurable
-maximum of tool calls, default 10).
+maximum of tool calls, default 25).
 
 ## Testing
 
-Run the multi-step planner tests with:
+Run the fast unit tests (no LLM calls):
 
 ```bash
-poetry run pytest global_chat/tests/test_planner_multistep.py -v -s
+poetry run pytest services/global_chat/tests/unit/ -q
 ```
 
-Run all tests for the service:
+Run all tests for the service (some hit live LLM APIs and cost tokens):
 
 ```bash
-poetry run pytest global_chat/tests/ -v -s
+poetry run pytest services/global_chat/tests/ -v -s
 ```

services/global_chat/config.yaml

@@ -11,4 +11,4 @@ planner:
   model: "claude-sonnet"
   max_tokens: 8192
   temperature: 1.0
-  max_tool_calls: 10
+  max_tool_calls: 25