[ccch1mneyyy]

Closes #339

What was wrong

The TUI activity observer only understood JSON text envelopes, but Native file tools and Native ops_result intentionally render compact model-facing prose. The sidebar counted returns while leaving successful reads, previews, refusals and operation errors business-unconfirmed. The optional real-host fixture also needed the explicit user-message source required by DSH 0.2.1.

The change

• Carry bounded outcome facts through the host's existing output.presentationMeta / committed meta.opsOutcome path. Model text and tool input/output schemas are unchanged.
• Read metadata for top-level Native results, keeping PTC child JSON handling separate. Host transport failure still cannot become confirmed business success.
• Exclude paths, content, items and error messages; keep only bounded status/error codes, effect facts, booleans and safe nonnegative counters.
• Add always-on producer/consumer regressions and update the real dsh-TUI fixture, paired panel docs and changelog. No new dependency or host registration changes.

Validation

DSH_OPS_REQUIRE_RUNTIME=1 node test/run.mjs tui-result tui-native tui-host
  3 suites passed, 0 skipped, 0 failed
  TUI result diagnostics: all 10 checks passed
  real dsh-TUI 0.16.1 worker/loop/UI integration exercised

node scripts/verify-gate.mjs --skip-install
  linux-x64: gate green
  126 suites passed; skipped: none
  manifest: 73 checks (minimum 59)
  native unit tests, runtime build/staging, packed release closure: passed

node scripts/check-file-size.mjs
  384 code files checked, all <=1000 lines; longest 924

git diff --check
  passed

Red/green: copying the new tui-result tests onto unmodified a9ff940
  2 of 10 checks fail: Native metadata classification and bounded metadata projection
  fixed tree: all 10 pass

The complete gate uses the host's real Landlock backend (bwrap probe deliberately unavailable on PATH) because main's unrelated private-/tmp reference-cache defect is fixed independently in #338. No unconfined substitute or permission relaxation is used. The targeted real-host Native suites also pass with the machine's ordinary bwrap-enabled PATH.

PR-triggered CI: https://github.com/T-Auto/dsh-ops/actions/runs/39477813622 — pull_request, aggregate CI required: SUCCESS (Windows x64, Linux x64, macOS arm64 and required supporting jobs passed). The first macOS job was abandoned before allocation (runner_id: 0, no steps). A targeted rerun of only that unexecuted job and its dependent aggregate succeeded on attempt 2; no failed code-test assertion was retried.

Environment

Arch Linux x64; Node 24.21.0; Rust 1.97.1; dsh-TUI 0.16.1; base main a9ff940. Local source checkout with rebuilt native worker.

Combined local validation

The submitted heads of #338, #340, #341, #342, #343, #344 and #345 were applied together in an isolated local verification worktree, including their combined changelog entries. With ordinary PATH selecting real bwrap 0.13.0 and real dsh-TUI 0.16.1 host/worker/renderer suites enabled, node scripts/verify-gate.mjs --skip-install passed all 11 release-gate steps: 129 suites passed, 0 skipped; 46 Rust unit tests passed; manifest 73/59. This Linux combined-tree evidence is separate from each PR's own cross-platform CI; no upstream merge or host installation was performed.