feat(fase2): wire BudgetMiddleware + retrieval tool onto the debate in run_project

src/portfolio_optimiser/run.py

@@ -31,9 +31,13 @@ from dataclasses import dataclass
 from agent_framework import BaseChatClient, SessionContext
 
 from portfolio_optimiser.backends import Profile, get_backend, resolve_model
-from portfolio_optimiser.budget import Budget, TokenMeter
+from portfolio_optimiser.budget import Budget, BudgetMiddleware, TokenMeter
 from portfolio_optimiser.contracts import load_contracts
-from portfolio_optimiser.datasource import chunk_dict_to_citation, retrieve_chunks
+from portfolio_optimiser.datasource import (
+    chunk_dict_to_citation,
+    make_retrieval_tool,
+    retrieve_chunks,
+)
 from portfolio_optimiser.generate import generate_via_llm
 from portfolio_optimiser.ir import SavingsProposal
 from portfolio_optimiser.provenance import ProvenanceStamp
@@ -118,9 +122,20 @@ async def run_project(
     context = "\n".join(c["snippet"] for c in chunks)
 
     # 4. Budget + maker-checker debate (round-capped; Layer-1 HITL optional).
+    # The shared meter is driven on the debate's chat calls by the BudgetMiddleware
+    # (the brief's named short-circuit mechanism); the retrieval tool exposes the
+    # citation-bearing data source to the agents (no longer string-stuffed out-of-band).
     meter = TokenMeter(Budget(max_tokens=max_tokens, max_rounds=max(max_rounds * 4, 4)))
     factory = client_factory if client_factory is not None else _default_factory(profile)
-    debate = fresh_workflow(factory, max_rounds=max_rounds, enable_layer1_hitl=enable_layer1_hitl)
+    budget_mw = BudgetMiddleware(meter)
+    retrieval_tool = make_retrieval_tool(docs_dir, top_k=top_k)
+    debate = fresh_workflow(
+        factory,
+        max_rounds=max_rounds,
+        enable_layer1_hitl=enable_layer1_hitl,
+        tools=[retrieval_tool],
+        middleware=[budget_mw],
+    )
     await debate.run(f"Find a cost-saving measure for {project.id}.\nContext:\n{context}")
 
     # 5. Structured candidate -> blocking validation; token bound = the meter in this loop.