feat: allow HITL to review materialized tool calls

[lorenzozanee]

Related Issues

• fixes Allow Human-in-the-Loop to review fully materialized tool calls #12060

Proposed Changes:

HITL currently shows only the LLM-produced arguments, while tools with inputs_from_state receive additional values from State at execution time. The reviewer therefore approves an incomplete view of the call.

This PR adds an opt-in mode to surface fully materialized calls before execution:

• BlockingConfirmationStrategy(include_state_inputs=True) and ConfirmationHook(include_state_inputs=True) materialize inputs_from_state via the same _inject_state_args used by tool_calling._run_tool
• default is False for backward compatibility
• hook flag propagates to strategies so a single ConfirmationHook(include_state_inputs=True) covers the wildcard * case
• State-typed parameters are shown as "<State>" to avoid dumping whole state
• serialization preserves the flag for both types and State missing keys are handled without error

The materialization happens at before_tool time from the current State snapshot, matching execution-time injection for the stable single-tool case. For multi-tool batches where an earlier tool writes a key read by a later one, the approved view reflects the state at approval time; the existing per-batch re-materialization at dispatch still determines the final execution value.

How did you test it?

• reproduction script with a tool using inputs_from_state={"selected_customer_id":"customer_id"} and State(selected_customer_id="cust_84319"): without the flag the UI sees {"reason":"duplicate"}, with the flag it sees {"reason":"duplicate","customer_id":"cust_84319"} and final_tool_params matches
• test/hooks/human_in_the_loop (99 passed) and test/components/agents/test_agent_hitl.py (7 passed) including new cases TestMaterializedStateInputs for strategy and hook propagation
• verified to_dict/from_dict round-trips for both BlockingConfirmationStrategy and ConfirmationHook
• hatch run fmt clean

Notes for the reviewer

Opt-in only, no behavior change by default. Follow-up provenance labeling (llm_arg vs state_injected) and redaction can build on this materialization path without changing the approval contract.

Checklist

• I have read the contributors guidelines and the code of conduct.
• I have updated the related issue with new insights and changes.
• I have added unit tests and updated the docstrings.
• I've used one of the conventional commit types for my PR title.
• I have added a release note file.
• I have run pre-commit hooks and fixed any issue.

[vercel]

@lorenzozanee is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[github-actions]

Hi @lorenzozanee, thanks for your interest in contributing to Haystack! 🙏

⚠️ You currently have 6 open pull requests in this repository (#12516, #12513, #12497, #12495, #12469 and this one). Our review capacity is limited, so please hold off opening more PRs until we've had a chance to review your first 2 open PRs. This helps us give each contribution the attention it deserves. Thank you!

This is an automated message to help us keep the review queue healthy.

[HaystackBot]

Hi @lorenzozanee, thanks a lot for your contribution! 🙏

We noticed that the Contributor License Agreement (CLA) check (license/cla) hasn't passed yet, so we've temporarily moved this PR to draft and paused the review assignment.

To get your PR reviewed, please sign the CLA via the link in the license/cla check below (or in the CLA bot comment). As soon as the check turns green, this PR will automatically be marked ready for review again and a reviewer will be re-assigned.