[AAASM-4840] 📝 (examples): Sync README dependency version floors to manifest pins

python/agno-tool-policy/README.md

@@ -35,7 +35,7 @@ Agno runs every function-tool call through a single chokepoint — `agno.tools.f
 | Python | >= 3.12 |
 | [uv](https://github.com/astral-sh/uv) | latest |
 | Agent Assembly Python SDK | >= 0.0.1rc6 |
-| Agno | >= 2.0.0 |
+| Agno | >= 2.7.2 |
 
 No API key or running gateway is required for the offline demo.
 

python/haystack-tool-policy/README.md

@@ -36,7 +36,7 @@ Haystack has a native adapter: `HaystackPatch` hooks `haystack.tools.Tool.invoke
 | Python | >= 3.12 |
 | [uv](https://github.com/astral-sh/uv) | latest |
 | Agent Assembly Python SDK | >= 0.0.1rc6 |
-| Haystack | >= 2.0.0, < 3.0 |
+| Haystack | >= 2.31.0, < 3.0 |
 
 No API key or running gateway is required for the offline demo — the tools are driven through a `ToolInvoker` with a hand-built `ToolCall`, so no LLM is involved.
 

python/pydantic-ai/README.md

@@ -36,7 +36,7 @@ Demonstrates how to integrate [Agent Assembly](https://github.com/ai-agent-assem
 
 No running Agent Assembly gateway is required for the offline demo.
 
-> **Version note** — the Agent Assembly Pydantic AI adapter installs a version-tolerant tool hook. On `pydantic-ai>=0.3.0` it patches `AbstractToolset.call_tool` (including concrete toolsets such as `FunctionToolset`); on `<0.3.0` it falls back to the internal `Tool._run`. `pyproject.toml` therefore requires `pydantic-ai>=0.3.0` with no upper pin — governance attaches across the `0.3.x`–`1.x` line.
+> **Version note** — the Agent Assembly Pydantic AI adapter installs a version-tolerant tool hook. On `pydantic-ai>=0.3.0` it patches `AbstractToolset.call_tool` (including concrete toolsets such as `FunctionToolset`); on `<0.3.0` it falls back to the internal `Tool._run`. `pyproject.toml` therefore requires `pydantic-ai>=2.9.0` with no upper pin.
 
 ## Setup
 
@@ -109,7 +109,7 @@ In production, `init_assembly()` auto-detects Pydantic AI and registers the adap
 |---|---|
 | `ModuleNotFoundError: agent_assembly` | Run `uv sync` first |
 | `ModuleNotFoundError: pydantic_ai` | Run `uv sync` — `pydantic-ai` is a required dependency |
-| Governance hooks do not fire | Ensure `pydantic-ai` resolves to `>=0.3.0` (or `<0.3.0` for the legacy `Tool._run` path) |
+| Governance hooks do not fire | Ensure `pydantic-ai` resolves to `>=2.9.0` (the manifest floor; the adapter's `AbstractToolset.call_tool` hook covers this range) |
 | `PolicyViolationError` in tests | Expected — the deny/pending policy rules are intentional |
 
 ## Links

python/semantic-kernel-tool-policy/README.md

@@ -37,7 +37,7 @@ Because Semantic Kernel has no native adapter, this example governs it the frame
 | Python | >= 3.12 |
 | [uv](https://github.com/astral-sh/uv) | latest |
 | Agent Assembly Python SDK | >= 0.0.1rc6 |
-| semantic-kernel | >= 1.36.0 |
+| semantic-kernel | >= 1.39.4 |
 
 No API key or running gateway is required for the offline demo — the tools are native (code) `KernelFunction`s, so no LLM connector is configured.
 

python/smolagents-tool-policy/README.md

@@ -41,7 +41,7 @@ This is not a mock: the tools are real `smolagents.Tool` subclasses and the gove
 | Python | >= 3.12 |
 | [uv](https://github.com/astral-sh/uv) | latest |
 | Agent Assembly Python SDK | >= 0.0.1rc6 |
-| smolagents | >= 1.0, < 2.0 |
+| smolagents | >= 1.26.0, < 2.0.0 |
 
 No running Agent Assembly gateway and no model/API credentials are required for the offline demo.