AbstractAgent 0.3.12__tar.gz → 0.3.13__tar.gz

{abstractagent-0.3.12 → abstractagent-0.3.13}/PKG-INFO

@@ -1,18 +1,18 @@
 Metadata-Version: 2.4
 Name: abstractagent
-Version: 0.3.12
+Version: 0.3.13
 Summary: Agent implementations using AbstractRuntime and AbstractCore
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: abstractcore[tools]>=2.13.38
-Requires-Dist: abstractruntime>=0.4.29
+Requires-Dist: abstractcore[tools]>=2.13.41
+Requires-Dist: abstractruntime>=0.4.32
 Provides-Extra: apple
-Requires-Dist: abstractcore[all-apple]>=2.13.38; extra == "apple"
-Requires-Dist: abstractruntime[apple]>=0.4.29; extra == "apple"
+Requires-Dist: abstractcore[all-apple]>=2.13.41; extra == "apple"
+Requires-Dist: abstractruntime[apple]>=0.4.32; extra == "apple"
 Provides-Extra: gpu
-Requires-Dist: abstractcore[all-gpu]>=2.13.38; extra == "gpu"
-Requires-Dist: abstractruntime[gpu]>=0.4.29; extra == "gpu"
+Requires-Dist: abstractcore[all-gpu]>=2.13.41; extra == "gpu"
+Requires-Dist: abstractruntime[gpu]>=0.4.32; extra == "gpu"
 Provides-Extra: dev
 Requires-Dist: pytest>=7.0; extra == "dev"
 Provides-Extra: test
@@ -45,6 +45,7 @@ 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)
 - Architecture (diagrams): [`docs/architecture.md`](docs/architecture.md)
 - Changelog: [`CHANGELOG.md`](CHANGELOG.md)
@@ -55,12 +56,16 @@ flowchart LR
 ## What you get
 
 - **ReAct**: tool-first Reason → Act → Observe loop
-- **CodeAct**: executes Python (tool call or fenced ` ```python``` ` blocks)
-- **MemAct**: memory-enhanced agent using runtime-owned Active Memory
+- **CodeAct**: executes Python (tool calls; on prompted-tools models also fenced ` ```python``` ` blocks)
+- **MemAct** (experimental): memory-enhanced agent using runtime-owned Active Memory — distinct architecture, less production mileage; import from `abstractagent.agents.memact` (deliberately not top-level)
 - **Durable runs**: pause/resume via `run_id` + runtime stores
-- **Tool control**: explicit tool bundles + per-run allowlists
-- **Generation controls**: temperature, seed, media policy, prompt-cache binding,
-  and Core `thinking` are normalized before LLM calls
+- **Tool control**: explicit tool bundles + per-run allowlists (unresolvable
+  grant names are recorded loudly, never dropped silently)
+- **Loop hooks**: listen/steer/capture on the running loop via `LoopHooks`
+  (see [`docs/hooks.md`](docs/hooks.md)); the flat `on_step` callback remains
+- **Generation controls**: temperature, seed, media policy, prompt-cache
+  identity, streaming, Core `thinking` and `speculation` (MTP) are normalized
+  before LLM calls and inherited by delegated children
 - **Observability**: durable ledger of LLM calls, tool calls, and waits
 
 Where this lives in code (source of truth):
@@ -98,6 +103,10 @@ 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).
+
 Note: the repository may be ahead of the latest published PyPI release. To verify what you installed:
 
 ```bash
@@ -109,12 +118,22 @@ python -c "import importlib.metadata as md; print(md.version('abstractagent'))"
 ```python
 from abstractagent import create_react_agent
 
-agent = create_react_agent(provider="ollama", model="qwen3:1.7b-q4_K_M")
+# provider/model resolve from your AbstractCore config defaults
+# (`abstractcore --config`) when omitted; pass them explicitly to pin.
+agent = create_react_agent(provider="ollama", model="qwen3:4b")
 agent.start("List the files in the current directory")
 state = agent.run_to_completion()
 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`.
+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)).
+
 ## Persistence (resume across restarts)
 
 By default, the factory helpers use an in-memory runtime store. For resume across process restarts,

abstractagent-0.3.12/src/abstractagent.egg-info/PKG-INFO → abstractagent-0.3.13/README.md

@@ -1,24 +1,3 @@
-Metadata-Version: 2.4
-Name: abstractagent
-Version: 0.3.12
-Summary: Agent implementations using AbstractRuntime and AbstractCore
-Requires-Python: >=3.10
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: abstractcore[tools]>=2.13.38
-Requires-Dist: abstractruntime>=0.4.29
-Provides-Extra: apple
-Requires-Dist: abstractcore[all-apple]>=2.13.38; extra == "apple"
-Requires-Dist: abstractruntime[apple]>=0.4.29; extra == "apple"
-Provides-Extra: gpu
-Requires-Dist: abstractcore[all-gpu]>=2.13.38; extra == "gpu"
-Requires-Dist: abstractruntime[gpu]>=0.4.29; extra == "gpu"
-Provides-Extra: dev
-Requires-Dist: pytest>=7.0; extra == "dev"
-Provides-Extra: test
-Requires-Dist: pytest>=7.0; extra == "test"
-Dynamic: license-file
-
 # AbstractAgent
 
 Agent patterns (ReAct / CodeAct / MemAct) built on **AbstractRuntime** (durable execution) and **AbstractCore** (tools + LLM integration).
@@ -45,6 +24,7 @@ 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)
 - Architecture (diagrams): [`docs/architecture.md`](docs/architecture.md)
 - Changelog: [`CHANGELOG.md`](CHANGELOG.md)
@@ -55,12 +35,16 @@ flowchart LR
 ## What you get
 
 - **ReAct**: tool-first Reason → Act → Observe loop
-- **CodeAct**: executes Python (tool call or fenced ` ```python``` ` blocks)
-- **MemAct**: memory-enhanced agent using runtime-owned Active Memory
+- **CodeAct**: executes Python (tool calls; on prompted-tools models also fenced ` ```python``` ` blocks)
+- **MemAct** (experimental): memory-enhanced agent using runtime-owned Active Memory — distinct architecture, less production mileage; import from `abstractagent.agents.memact` (deliberately not top-level)
 - **Durable runs**: pause/resume via `run_id` + runtime stores
-- **Tool control**: explicit tool bundles + per-run allowlists
-- **Generation controls**: temperature, seed, media policy, prompt-cache binding,
-  and Core `thinking` are normalized before LLM calls
+- **Tool control**: explicit tool bundles + per-run allowlists (unresolvable
+  grant names are recorded loudly, never dropped silently)
+- **Loop hooks**: listen/steer/capture on the running loop via `LoopHooks`
+  (see [`docs/hooks.md`](docs/hooks.md)); the flat `on_step` callback remains
+- **Generation controls**: temperature, seed, media policy, prompt-cache
+  identity, streaming, Core `thinking` and `speculation` (MTP) are normalized
+  before LLM calls and inherited by delegated children
 - **Observability**: durable ledger of LLM calls, tool calls, and waits
 
 Where this lives in code (source of truth):
@@ -98,6 +82,10 @@ 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).
+
 Note: the repository may be ahead of the latest published PyPI release. To verify what you installed:
 
 ```bash
@@ -109,12 +97,22 @@ python -c "import importlib.metadata as md; print(md.version('abstractagent'))"
 ```python
 from abstractagent import create_react_agent
 
-agent = create_react_agent(provider="ollama", model="qwen3:1.7b-q4_K_M")
+# provider/model resolve from your AbstractCore config defaults
+# (`abstractcore --config`) when omitted; pass them explicitly to pin.
+agent = create_react_agent(provider="ollama", model="qwen3:4b")
 agent.start("List the files in the current directory")
 state = agent.run_to_completion()
 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`.
+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)).
+
 ## Persistence (resume across restarts)
 
 By default, the factory helpers use an in-memory runtime store. For resume across process restarts,

{abstractagent-0.3.12 → abstractagent-0.3.13}/pyproject.toml

@@ -4,23 +4,23 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "abstractagent"
-version = "0.3.12"
+version = "0.3.13"
 description = "Agent implementations using AbstractRuntime and AbstractCore"
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
-    "abstractcore[tools]>=2.13.38",
-    "abstractruntime>=0.4.29",
+    "abstractcore[tools]>=2.13.41",
+    "abstractruntime>=0.4.32",
 ]
 
 [project.optional-dependencies]
 apple = [
-    "abstractcore[all-apple]>=2.13.38",
-    "abstractruntime[apple]>=0.4.29",
+    "abstractcore[all-apple]>=2.13.41",
+    "abstractruntime[apple]>=0.4.32",
 ]
 gpu = [
-    "abstractcore[all-gpu]>=2.13.38",
-    "abstractruntime[gpu]>=0.4.29",
+    "abstractcore[all-gpu]>=2.13.41",
+    "abstractruntime[gpu]>=0.4.32",
 ]
 dev = [
     "pytest>=7.0",

{abstractagent-0.3.12 → abstractagent-0.3.13}/src/abstractagent/__init__.py

@@ -1,5 +1,12 @@
 """AbstractAgent - Agent implementations using AbstractRuntime and AbstractCore."""
 
+from .adapters.loop_hooks import DEFAULT_EVENT_MAP, HookEvent, LoopHooks
+from .adapters.native_loop_registry import (
+    audit_native_loop_manifest,
+    build_native_loop_manifest,
+    materialize_native_loop_spec,
+    pack_native_loop_bundle,
+)
 from .agents import (
     BaseAgent,
     ReactAgent,
@@ -26,6 +33,14 @@ from .tools import (
 __all__ = [
     # Base class for custom agents
     "BaseAgent",
+    # First-class loop hooks (listen + steer + capture)
+    "DEFAULT_EVENT_MAP",
+    "HookEvent",
+    "LoopHooks",
+    "audit_native_loop_manifest",
+    "materialize_native_loop_spec",
+    "build_native_loop_manifest",
+    "pack_native_loop_bundle",
     # ReAct agent
     "ReactAgent",
     "create_react_workflow",

abstractagent-0.3.13/src/abstractagent/adapters/__init__.py

@@ -0,0 +1,21 @@
+"""Runtime adapters for agent logic."""
+
+from .codeact_runtime import create_codeact_workflow
+from .native_loop_registry import (
+    audit_native_loop_manifest,
+    build_native_loop_manifest,
+    materialize_native_loop_spec,
+    pack_native_loop_bundle,
+)
+from .react_runtime import create_react_workflow
+from .memact_runtime import create_memact_workflow
+
+__all__ = [
+    "create_react_workflow",
+    "create_codeact_workflow",
+    "create_memact_workflow",
+    "audit_native_loop_manifest",
+    "materialize_native_loop_spec",
+    "build_native_loop_manifest",
+    "pack_native_loop_bundle",
+]