AbstractAgent 0.3.13__tar.gz → 0.3.14__tar.gz

{abstractagent-0.3.13/src/abstractagent.egg-info → abstractagent-0.3.14}/PKG-INFO

@@ -1,18 +1,18 @@
 Metadata-Version: 2.4
 Name: abstractagent
-Version: 0.3.13
+Version: 0.3.14
 Summary: Agent implementations using AbstractRuntime and AbstractCore
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: abstractcore[tools]>=2.13.41
-Requires-Dist: abstractruntime>=0.4.32
+Requires-Dist: abstractruntime>=0.5.0
 Provides-Extra: apple
 Requires-Dist: abstractcore[all-apple]>=2.13.41; extra == "apple"
-Requires-Dist: abstractruntime[apple]>=0.4.32; extra == "apple"
+Requires-Dist: abstractruntime[apple]>=0.5.0; extra == "apple"
 Provides-Extra: gpu
 Requires-Dist: abstractcore[all-gpu]>=2.13.41; extra == "gpu"
-Requires-Dist: abstractruntime[gpu]>=0.4.32; extra == "gpu"
+Requires-Dist: abstractruntime[gpu]>=0.5.0; extra == "gpu"
 Provides-Extra: dev
 Requires-Dist: pytest>=7.0; extra == "dev"
 Provides-Extra: test
@@ -46,12 +46,14 @@ flowchart LR
 - Getting started: [`docs/getting-started.md`](docs/getting-started.md)
 - API reference: [`docs/api.md`](docs/api.md)
 - Loop hooks (listen/steer/capture): [`docs/hooks.md`](docs/hooks.md)
-- FAQ / troubleshooting: [`docs/faq.md`](docs/faq.md)
+- FAQ: [`docs/faq.md`](docs/faq.md)
+- Troubleshooting: [`docs/troubleshooting.md`](docs/troubleshooting.md)
 - Architecture (diagrams): [`docs/architecture.md`](docs/architecture.md)
 - Changelog: [`CHANGELOG.md`](CHANGELOG.md)
 - Contributing: [`CONTRIBUTING.md`](CONTRIBUTING.md)
 - Security: [`SECURITY.md`](SECURITY.md)
 - Acknowledgements: [`ACKNOWLEDMENTS.md`](ACKNOWLEDMENTS.md)
+- Code of conduct: [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md)
 
 ## What you get
 
@@ -103,9 +105,10 @@ Native Python hardware profile cascades are available for deployment manifests:
 AbstractCore and AbstractRuntime profiles; AbstractAgent itself remains
 provider/runtime agnostic.
 
-AbstractAgent 0.3.13 requires `abstractcore[tools]>=2.13.41` and
-`AbstractRuntime>=0.4.32` (the runtime release that provides turn grounding,
-in-flight cancellation and speculation inheritance).
+AbstractAgent 0.3.14 requires `abstractcore[tools]>=2.13.41` and
+`AbstractRuntime>=0.5.0` (the runtime release that provides live token
+streaming, on top of turn grounding, in-flight cancellation and speculation
+inheritance). AbstractRuntime 0.5.0 itself requires AbstractCore 2.16.0 or newer.
 
 Note: the repository may be ahead of the latest published PyPI release. To verify what you installed:
 
@@ -129,8 +132,7 @@ print(state.output["answer"])
 Tip: these loops send the full transcript plus ~19 tool schemas every cycle —
 prefer a tool-capable model. On Ollama, raise the context window to the model's
 maximum available context or the server silently truncates from the oldest
-content first: per-call `llm_kwargs={"num_ctx": <model max>}` (forwarded since
-2026-07-13) or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
+content first: per-call `llm_kwargs={"num_ctx": <model max>}` or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
 House rule: maximum available context unless you explicitly choose otherwise —
 a fixed lower number is a hidden ceiling (see [`docs/faq.md`](docs/faq.md)).
 

{abstractagent-0.3.13 → abstractagent-0.3.14}/README.md

@@ -25,12 +25,14 @@ flowchart LR
 - Getting started: [`docs/getting-started.md`](docs/getting-started.md)
 - API reference: [`docs/api.md`](docs/api.md)
 - Loop hooks (listen/steer/capture): [`docs/hooks.md`](docs/hooks.md)
-- FAQ / troubleshooting: [`docs/faq.md`](docs/faq.md)
+- FAQ: [`docs/faq.md`](docs/faq.md)
+- Troubleshooting: [`docs/troubleshooting.md`](docs/troubleshooting.md)
 - Architecture (diagrams): [`docs/architecture.md`](docs/architecture.md)
 - Changelog: [`CHANGELOG.md`](CHANGELOG.md)
 - Contributing: [`CONTRIBUTING.md`](CONTRIBUTING.md)
 - Security: [`SECURITY.md`](SECURITY.md)
 - Acknowledgements: [`ACKNOWLEDMENTS.md`](ACKNOWLEDMENTS.md)
+- Code of conduct: [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md)
 
 ## What you get
 
@@ -82,9 +84,10 @@ Native Python hardware profile cascades are available for deployment manifests:
 AbstractCore and AbstractRuntime profiles; AbstractAgent itself remains
 provider/runtime agnostic.
 
-AbstractAgent 0.3.13 requires `abstractcore[tools]>=2.13.41` and
-`AbstractRuntime>=0.4.32` (the runtime release that provides turn grounding,
-in-flight cancellation and speculation inheritance).
+AbstractAgent 0.3.14 requires `abstractcore[tools]>=2.13.41` and
+`AbstractRuntime>=0.5.0` (the runtime release that provides live token
+streaming, on top of turn grounding, in-flight cancellation and speculation
+inheritance). AbstractRuntime 0.5.0 itself requires AbstractCore 2.16.0 or newer.
 
 Note: the repository may be ahead of the latest published PyPI release. To verify what you installed:
 
@@ -108,8 +111,7 @@ print(state.output["answer"])
 Tip: these loops send the full transcript plus ~19 tool schemas every cycle —
 prefer a tool-capable model. On Ollama, raise the context window to the model's
 maximum available context or the server silently truncates from the oldest
-content first: per-call `llm_kwargs={"num_ctx": <model max>}` (forwarded since
-2026-07-13) or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
+content first: per-call `llm_kwargs={"num_ctx": <model max>}` or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
 House rule: maximum available context unless you explicitly choose otherwise —
 a fixed lower number is a hidden ceiling (see [`docs/faq.md`](docs/faq.md)).
 

{abstractagent-0.3.13 → abstractagent-0.3.14}/pyproject.toml

@@ -4,23 +4,23 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "abstractagent"
-version = "0.3.13"
+version = "0.3.14"
 description = "Agent implementations using AbstractRuntime and AbstractCore"
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
     "abstractcore[tools]>=2.13.41",
-    "abstractruntime>=0.4.32",
+    "abstractruntime>=0.5.0",
 ]
 
 [project.optional-dependencies]
 apple = [
     "abstractcore[all-apple]>=2.13.41",
-    "abstractruntime[apple]>=0.4.32",
+    "abstractruntime[apple]>=0.5.0",
 ]
 gpu = [
     "abstractcore[all-gpu]>=2.13.41",
-    "abstractruntime[gpu]>=0.4.32",
+    "abstractruntime[gpu]>=0.5.0",
 ]
 dev = [
     "pytest>=7.0",

{abstractagent-0.3.13 → abstractagent-0.3.14}/src/abstractagent/adapters/codeact_runtime.py

@@ -1094,6 +1094,10 @@ def create_codeact_workflow(
                     # that silently reverts to provider defaults.
                     "thinking",
                     "speculation",
+                    # Live token streaming (2026-09-26): a delegated child
+                    # streams its LLM calls to the same live view as the
+                    # parent; False (explicit off) is inherited too.
+                    "stream",
                     "max_output_tokens",
                     "tool_prompt_examples",
                     # Approval policy inherits monotonically (tool-tiers adversary

{abstractagent-0.3.13 → abstractagent-0.3.14}/src/abstractagent/adapters/memact_runtime.py

@@ -915,6 +915,10 @@ def create_memact_workflow(
                     # that silently reverts to provider defaults.
                     "thinking",
                     "speculation",
+                    # Live token streaming (2026-09-26): a delegated child
+                    # streams its LLM calls to the same live view as the
+                    # parent; False (explicit off) is inherited too.
+                    "stream",
                     "max_output_tokens",
                     "tool_prompt_examples",
                     # Approval policy inherits monotonically (tool-tiers adversary

{abstractagent-0.3.13 → abstractagent-0.3.14}/src/abstractagent/adapters/react_runtime.py

@@ -2871,6 +2871,10 @@ def create_react_workflow(
                     # that silently reverts to provider defaults.
                     "thinking",
                     "speculation",
+                    # Live token streaming (2026-09-26): a delegated child
+                    # streams its LLM calls to the same live view as the
+                    # parent; False (explicit off) is inherited too.
+                    "stream",
                     "max_output_tokens",
                     "tool_prompt_examples",
                     # Approval policy inherits MONOTONICALLY (tool-tiers adversary

{abstractagent-0.3.13 → abstractagent-0.3.14/src/abstractagent.egg-info}/PKG-INFO

@@ -1,18 +1,18 @@
 Metadata-Version: 2.4
 Name: abstractagent
-Version: 0.3.13
+Version: 0.3.14
 Summary: Agent implementations using AbstractRuntime and AbstractCore
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: abstractcore[tools]>=2.13.41
-Requires-Dist: abstractruntime>=0.4.32
+Requires-Dist: abstractruntime>=0.5.0
 Provides-Extra: apple
 Requires-Dist: abstractcore[all-apple]>=2.13.41; extra == "apple"
-Requires-Dist: abstractruntime[apple]>=0.4.32; extra == "apple"
+Requires-Dist: abstractruntime[apple]>=0.5.0; extra == "apple"
 Provides-Extra: gpu
 Requires-Dist: abstractcore[all-gpu]>=2.13.41; extra == "gpu"
-Requires-Dist: abstractruntime[gpu]>=0.4.32; extra == "gpu"
+Requires-Dist: abstractruntime[gpu]>=0.5.0; extra == "gpu"
 Provides-Extra: dev
 Requires-Dist: pytest>=7.0; extra == "dev"
 Provides-Extra: test
@@ -46,12 +46,14 @@ flowchart LR
 - Getting started: [`docs/getting-started.md`](docs/getting-started.md)
 - API reference: [`docs/api.md`](docs/api.md)
 - Loop hooks (listen/steer/capture): [`docs/hooks.md`](docs/hooks.md)
-- FAQ / troubleshooting: [`docs/faq.md`](docs/faq.md)
+- FAQ: [`docs/faq.md`](docs/faq.md)
+- Troubleshooting: [`docs/troubleshooting.md`](docs/troubleshooting.md)
 - Architecture (diagrams): [`docs/architecture.md`](docs/architecture.md)
 - Changelog: [`CHANGELOG.md`](CHANGELOG.md)
 - Contributing: [`CONTRIBUTING.md`](CONTRIBUTING.md)
 - Security: [`SECURITY.md`](SECURITY.md)
 - Acknowledgements: [`ACKNOWLEDMENTS.md`](ACKNOWLEDMENTS.md)
+- Code of conduct: [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md)
 
 ## What you get
 
@@ -103,9 +105,10 @@ Native Python hardware profile cascades are available for deployment manifests:
 AbstractCore and AbstractRuntime profiles; AbstractAgent itself remains
 provider/runtime agnostic.
 
-AbstractAgent 0.3.13 requires `abstractcore[tools]>=2.13.41` and
-`AbstractRuntime>=0.4.32` (the runtime release that provides turn grounding,
-in-flight cancellation and speculation inheritance).
+AbstractAgent 0.3.14 requires `abstractcore[tools]>=2.13.41` and
+`AbstractRuntime>=0.5.0` (the runtime release that provides live token
+streaming, on top of turn grounding, in-flight cancellation and speculation
+inheritance). AbstractRuntime 0.5.0 itself requires AbstractCore 2.16.0 or newer.
 
 Note: the repository may be ahead of the latest published PyPI release. To verify what you installed:
 
@@ -129,8 +132,7 @@ print(state.output["answer"])
 Tip: these loops send the full transcript plus ~19 tool schemas every cycle —
 prefer a tool-capable model. On Ollama, raise the context window to the model's
 maximum available context or the server silently truncates from the oldest
-content first: per-call `llm_kwargs={"num_ctx": <model max>}` (forwarded since
-2026-07-13) or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
+content first: per-call `llm_kwargs={"num_ctx": <model max>}` or server-side `OLLAMA_CONTEXT_LENGTH=<model max> ollama serve`.
 House rule: maximum available context unless you explicitly choose otherwise —
 a fixed lower number is a hidden ceiling (see [`docs/faq.md`](docs/faq.md)).
 

{abstractagent-0.3.13 → abstractagent-0.3.14}/src/abstractagent.egg-info/SOURCES.txt

@@ -51,6 +51,7 @@ tests/test_active_context_policy_adoption.py
 tests/test_adversary_hardening_2026_07_13.py
 tests/test_agent_plan_review_modes.py
 tests/test_agent_speculation_inheritance.py
+tests/test_agent_stream_inheritance.py
 tests/test_agents_integration_llm.py
 tests/test_base_agent_trace_accessors.py
 tests/test_batching_prompt_rule.py

{abstractagent-0.3.13 → abstractagent-0.3.14}/src/abstractagent.egg-info/requires.txt

@@ -1,16 +1,16 @@
 abstractcore[tools]>=2.13.41
-abstractruntime>=0.4.32
+abstractruntime>=0.5.0
 
 [apple]
 abstractcore[all-apple]>=2.13.41
-abstractruntime[apple]>=0.4.32
+abstractruntime[apple]>=0.5.0
 
 [dev]
 pytest>=7.0
 
 [gpu]
 abstractcore[all-gpu]>=2.13.41
-abstractruntime[gpu]>=0.4.32
+abstractruntime[gpu]>=0.5.0
 
 [test]
 pytest>=7.0

abstractagent-0.3.14/tests/test_agent_stream_inheritance.py

@@ -0,0 +1,56 @@
+"""Delegated sub-agents inherit the run's live token-streaming switch.
+
+`_runtime.stream: true` asks the runtime to stream every LLM call of the run
+to the host's live view. A `delegate_agent` child is started with an explicit
+`_runtime` rider list; without `stream` in it, the child's calls would silently
+stop streaming mid-tree. False (explicit off) is inherited too; an absent value
+stays absent.
+"""
+from types import SimpleNamespace
+
+import pytest
+
+from abstractagent.adapters.codeact_runtime import create_codeact_workflow
+from abstractagent.adapters.memact_runtime import create_memact_workflow
+from abstractagent.adapters.react_runtime import create_react_workflow
+from abstractagent.logic.builtins import DELEGATE_AGENT_TOOL
+from abstractagent.logic.codeact import CodeActLogic
+from abstractagent.logic.memact import MemActLogic
+from abstractagent.logic.react import ReActLogic
+from abstractruntime import EffectType, RunState, RunStatus
+
+
+@pytest.mark.parametrize(
+    "factory,logic",
+    [
+        (create_react_workflow, ReActLogic),
+        (create_codeact_workflow, CodeActLogic),
+        (create_memact_workflow, MemActLogic),
+    ],
+)
+@pytest.mark.parametrize("stream", [True, False, None])
+def test_every_delegate_loop_inherits_stream(factory, logic, stream):
+    workflow = factory(logic=logic(tools=[DELEGATE_AGENT_TOOL]), workflow_id="wf")
+    runtime_ns = {"inbox": [], "provider": "stub", "model": "stub", "allowed_tools": ["delegate_agent"]}
+    if stream is not None:
+        runtime_ns["stream"] = stream
+    run = RunState(
+        run_id="parent",
+        workflow_id="wf",
+        status=RunStatus.RUNNING,
+        current_node="act",
+        vars={
+            "context": {"task": "test", "messages": []},
+            "scratchpad": {"iteration": 1, "max_iterations": 20, "cycles": []},
+            "_runtime": runtime_ns,
+            "_temp": {"pending_tool_calls": [{"name": "delegate_agent", "arguments": {"task": "subtask"}, "call_id": "d1"}]},
+            "_limits": {"max_iterations": 20},
+        },
+    )
+    plan = workflow.get_node("act")(run, SimpleNamespace(now_iso=lambda: "2026-09-26T00:00:00Z"))
+    assert plan.effect is not None and plan.effect.type == EffectType.START_SUBWORKFLOW
+    child = plan.effect.payload["vars"]["_runtime"]
+    if stream is None:
+        assert "stream" not in child
+    else:
+        assert child["stream"] is stream

{abstractagent-0.3.13 → abstractagent-0.3.14}/tests/test_packaging_profiles.py

@@ -19,7 +19,7 @@ def test_base_dependency_floors_match_gateway_alignment() -> None:
     deps = set(_project()["dependencies"])
 
     assert "abstractcore[tools]>=2.13.41" in deps
-    assert "abstractruntime>=0.4.32" in deps
+    assert "abstractruntime>=0.5.0" in deps
 
 
 def test_hardware_profile_extras_are_core_runtime_cascades() -> None:
@@ -27,11 +27,11 @@ def test_hardware_profile_extras_are_core_runtime_cascades() -> None:
 
     assert set(extras["apple"]) == {
         "abstractcore[all-apple]>=2.13.41",
-        "abstractruntime[apple]>=0.4.32",
+        "abstractruntime[apple]>=0.5.0",
     }
     assert set(extras["gpu"]) == {
         "abstractcore[all-gpu]>=2.13.41",
-        "abstractruntime[gpu]>=0.4.32",
+        "abstractruntime[gpu]>=0.5.0",
     }
     assert "all-apple" not in extras
     assert "all-gpu" not in extras