Move configuration to a YAML file, and make the tuning knobs configurable

The thresholds Part H says to tune on the real inference hardware were
hardcoded, so every adjustment would have been a PR. They now live in a
`tuning:` section of a YAML config file, which also replaces flat
environment variables as the home for structure (#32).

- The file (NISSE_CONFIG, default ./nisse-dev.yaml) carries structure
  and holds no secrets: credentials are ${VAR} references resolved from
  the environment, and a reference to an unset variable is an error
- The environment still wins: every existing NISSE_* variable overrides
  its key, and the file is optional, so .env and CI work unchanged
- Validation stays fail-fast and reports every problem at once; unknown
  keys are errors, so a misspelt knob fails at startup instead of
  silently doing nothing
- The config file joins the state store and .env in the check that
  refuses to start when they sit inside the workspace clone
- Tuning covers the executor's turn and failure limits, the consult
  budget, the delegation caps, the command timeout and the tool output
  limit, which Workspace now takes as a setting; the defaults are the
  values the agent already ran with
- llm.max_output_tokens reaches both LLM clients from the config

docs/Tuning-Guide.md explains the knobs for someone learning local
model tuning: how weights, quantization and the KV cache spend VRAM,
what each knob does and what goes wrong at either extreme, and an order
to tune in. nisse-dev.example.yaml documents every key and is checked
by a test so it cannot drift.

Closes #32

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

nisse-dev/src/nisse_dev/agent.py

@@ -16,7 +16,7 @@ from nisse_dev.cancel import Cancel
 from nisse_dev.catalog import ModelCatalog, select_client
 from nisse_dev.config import AgentConfig
 from nisse_dev.control import Orchestrator
-from nisse_dev.delegation import Delegation
+from nisse_dev.delegation import Caps, Delegation
 from nisse_dev.executor import LLMExecutor
 from nisse_dev.expert import ExpertDesk
 from nisse_dev.git import GitRepo, authenticated_url
@@ -149,18 +149,29 @@ def build_agent(
     account = gitea.whoami()
 
     git_repo = prepare_workspace(config, account)
-    workspace = Workspace(config.workspace)
+    tuning = config.tuning
+    workspace = Workspace(
+        config.workspace,
+        timeout_s=tuning.tools.command_timeout_s,
+        max_output_chars=tuning.tools.max_output_chars,
+    )
 
-    normalized = LiteLLMClient(config.llm_base_url, config.llm_token)
+    normalized = LiteLLMClient(
+        config.llm_base_url, config.llm_token, max_tokens=config.max_output_tokens
+    )
     # Delegations prefer the vendor-native route; fidelity is cheap
     # elsewhere (A5). Which route a given model gets is decided by
     # select_client, not by where the caller happens to sit.
-    passthrough = AnthropicPassthroughClient(config.llm_base_url, config.llm_token)
+    passthrough = AnthropicPassthroughClient(
+        config.llm_base_url, config.llm_token, max_tokens=config.max_output_tokens
+    )
 
     transport = MatrixTransport(config.matrix)
     announce = notify or transport.send_message
 
-    expert = ExpertDesk(normalized, store, config.roles)
+    expert = ExpertDesk(
+        normalized, store, config.roles, budget=tuning.expert.consult_budget
+    )
     delegate_model = config.roles.resolve("delegate")
     delegation = Delegation(
         select_client(delegate_model, normalized, passthrough, catalog),
@@ -169,6 +180,11 @@ def build_agent(
         git_repo,
         model=delegate_model,
         backing_model=catalog.facts(delegate_model).backing,
+        caps=Caps(
+            turns=tuning.delegation.max_turns,
+            tokens=tuning.delegation.max_tokens,
+            wall_clock_s=tuning.delegation.wall_clock_s,
+        ),
         cancel=cancel,
     )
     executor = LLMExecutor(
@@ -180,6 +196,8 @@ def build_agent(
         delegate=delegation,
         notify=announce,
         cancel=cancel,
+        max_turns=tuning.executor.max_turns,
+        failure_limit=tuning.executor.failure_limit,
     )
     publisher = MilestonePublisher(
         store,