fix(daemon): carry projectless refusals as structured content

[ScriptedAlchemy]

This fixes the #2335 follow-up. A refusal answered on the daemon's projectless route (for example tracedecay serve outside a project, or tracedecay_project_list through the profile owner) put the typed problem record beside content, at the top level of the tools/call result. CallToolResult has no extension members, so MCP clients received that refusal only as prose. Project connections already move the record into structuredContent, in the rmcp call_tool adapter.

There is now one renderer, tracedecay_mcp::tool_errors::structure_tool_problem. The rmcp call_tool adapter and the projectless tools/call response both call it, so every host-facing tools/call refusal has the same isError shape: {content, isError: true, structuredContent: {problem}}. The inline copy of the move in rmcp.rs is deleted.

Fail-before / pass-after

Test: daemon::tests::socket::projectless_project_list_reads_the_empty_profile_registry, a real projectless socket connection. It now asserts the refusal record under structuredContent.problem and asserts that result.problem is absent.

On origin/master source (b836d91) with only this test changed:

panicked at crates/tracedecay/src/daemon/tests/socket.rs:813:5:
  left: (Number(9), Bool(true), Some(Object {"cancellation_stage": Null, "code": String("application.surface.invalid_request"), ...
 right: (Number(9), Bool(true), None, String("invalid_request"), String("application.surface.invalid_request"), String("invalid arguments for tracedecay_project_list: unknown field `limt`, expected `limit`"))
test result: FAILED. 0 passed; 1 failed

With this change: ok.

Runtime journey

Setup: debug tracedecay built from this branch, isolated HOME/profile, one daemon under systemd-run --user --scope -p MemoryMax=6G -p MemorySwapMax=1G, stopped afterwards.

tracedecay serve (cwd outside any project):
tools/list: 230 tools
tracedecay_project_list {"limt": 5}: isError=True 3ms
     structuredContent.problem.code="application.surface.invalid_request" top-level problem.code=null keys=['content', 'isError', 'structuredContent']
tracedecay_project_list {"format": "json", "limit": 1}: isError=None 8ms

Before this change, the same call answered keys=['content', 'isError', 'problem'] with structuredContent absent. That was observed on b836d91 with the same driver.

Checks

• tracedecay-mcp lib: 387 passed.
• Root --lib -- daemon:: mcp::: 735 passed, 2 failed, neither in the code this PR touches:
  • runtime_identity::concurrent_same_identity_worktrees… is master-red (test: three suites still expect JSON-RPC errors for owner refusals now rendered as isError #2344).
  • http_application_tests::daemon_http_shutdown_releases_loopback_listener failed rebinding a loopback port (AddrInUse) under the shared host's load.
• cargo clippy -p tracedecay -p tracedecay-mcp --all-targets -D warnings, with and without tracedecay/test-transport,test-helpers: clean. cargo fmt --all -- --check: clean.

[changeset-bot]

⚠️ No Changeset found

Latest commit: f1059c5

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-27T17:04:40.388599Z	f1059c5	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.