fix(runtime_provision): use team-template risk_profile, not hardcoded toolfree

The provisioner was hardcoding risk_profile=toolfree for every claw,
which the ZeroClaw config explicitly configures to EXCLUDE every
usable tool (shell, file_read, file_write, http_request, browser).
Result: coder/tester/committer claws had zero tools and produced text
in the context window with no ability to actually write files or run
tests — exactly what the last mission summary showed.

Fixes:
- provision_claw now takes risk_profile: &str, passed through from
  the team template (development teams already had coding_readwrite,
  which now actually gets applied).
- Research team templates updated from toolfree → research_readonly
  (file_read) and papers_research → research_web_readonly
  (file_read + web_search + web_fetch). Applied to both the on-disk
  TOML files and the live DB rows.
- Added RuntimeProvisioner::default_risk_profile_for_role for
  auto-provision code paths that lack a template context — picks
  coding_readwrite for coder-like roles, research_readonly otherwise.
- Split rebind_model out of provision_claw so the model-change UI
  path doesnt inadvertently clobber the existing risk_profile.

Templates DB fixup for missions launched pre-deploy is already
applied via manual UPDATE.

crates/cm-api/src/routes/claws.rs

@@ -994,7 +994,7 @@ pub async fn set_model(
     //    model on their next turn. provision_claw overwrites
     //    agents.<alias>.model_provider on the shared ZeroClaw config.
     if let Some(provisioner) = crate::runtime_provision::RuntimeProvisioner::from_env() {
-        if let Err(e) = provisioner.provision_claw(id.as_uuid(), model).await {
+        if let Err(e) = provisioner.rebind_model(id.as_uuid(), model).await {
             eprintln!("set_model({id}): runtime rebind failed: {e}");
         }
     }