fix(miner-ui): resolve governor pause/resume chat intents to the registered actions

[tryeverything24]

Closes #8670

What

apps/loopover-miner-ui's governor pause/resume chat-action family was fully built and unit-tested — runGovernorChatAction (dispatch wrapper), GovernorChatActionResult (result renderer), chat-governor-action-copy.ts (Ledgers-verbatim copy) — but dead in the product: conversation.tsx's handleSubmit only resolved portfolio-queue text, so an operator typing "pause the governor" into the chat rail was silently answered by the generic read-only streaming assistant. runGovernorChatAction had zero non-test callers and GovernorChatActionResult was never imported outside its own test, exactly as the issue's repo-wide grep found.

How

All four Deliverables, in this one PR:

• New text resolver — src/lib/chat-governor-resolve.ts (resolveGovernorChatAction) mirrors chat-portfolio-queue-resolve.ts's pattern: pure (no fetch, no dispatch), word-bounded pause/halt/suspend vs resume/unpause intent detection that must co-occur with the word "governor", an optional pause reason from a trailing because … / reason: … clause (mapped to the { reason } params shape isGovernorPauseChatParams validates), and an explicit unresolvable result for empty/ambiguous/verb-less text so a governor question still falls through to the assistant and nothing is ever best-guess dispatched.
• handleSubmit wiring — conversation.tsx resolves governor intent right after the portfolio-queue check and, on a match, performs the same idempotent registerGovernorChatActions() registration the Vite dev-server entry does, then dispatches through runGovernorChatAction (flag-gated dispatchChatAction + governorGatedHandler, never a direct handler call). Unresolved text streams exactly as before.
• GovernorChatActionResult rendered in the message stream — the in-flight round-trip renders the component's pending state (role="status", "Updating governor…") in the same viewport slot the live streaming render uses, and the resolved turn is committed to the message list with the result attached, where MessageBubble now renders it through GovernorChatActionResult (error → role="alert", success → the Ledgers-verbatim "Paused since …"/"Not paused" copy). Non-executed dispatches (flag off / gated) and thrown dispatches surface as inline system notes, mirroring the existing failure convention.
• End-to-end RTL test — chat-conversation.test.tsx gains a 7-test #8670 describe that renders ChatConversation, types governor-intent text into the real composer, and drives the REAL registration + dispatch pipeline (createChatActionRegistry() + injected pauseGovernor/resumeGovernor clients only): asserts the pause action actually fires, the because-clause reaches pauseGovernor("release traffic spiked") as the structured reason, resume fires, GovernorChatActionResult's copy appears in the message list, pending copy shows while outstanding, non-executed/thrown dispatches surface, and streamChat is never called for action turns (0 calls) while ordinary governor questions still stream. Resolver branches are pinned in chat-governor-resolve.test.ts.

Before / after (RED proof)

The new end-to-end tests on unfixed main (fix stashed, tests kept) fail exactly as the issue describes — governor-intent text falls through to the generic assistant and no action ever fires:

× END-TO-END: typing "pause the governor" fires the pause action and renders GovernorChatActionResult copy
  AssertionError: expected "vi.fn()" to be called 1 times, but got 0 times
× typing "resume the governor" fires the resume action and renders the not-paused copy
  AssertionError: expected "vi.fn()" to be called 1 times, but got 0 times
(6 of the 7 new tests fail on main; the falls-through-to-assistant test passes, pinning the unchanged path)

With the fix, the full app suite:

Test Files  38 passed (38)
     Tests  407 passed (407)

Local vitest coverage thresholds (85/85/75/85, enforced in CI for this app) pass at 91.81% statements / 89.64% branches / 87.55% functions / 94.36% lines; every new/changed file is at 100% statements (chat-governor-resolve.ts 100% across all four metrics). tsc --noEmit, eslint and prettier --check on the changed files, and git diff --check are all green.

UI Evidence

The miner-ui chat rail on the Overview route, served by vite dev in the app's own demo mode (VITE_DEMO_MODE=1, sample data, no live backend), captured per viewport × theme below. The read-only POST /api/chat stream is answered at the network boundary by a fixture (the generic grounded assistant reply); the governor path needs no fixture at all — in demo mode pauseGovernor performs the real demo-state write, so the After column is the actual resolver → registration → flag-gated dispatch → GovernorChatActionResult path executing in the browser. Look at the reply to "pause the governor": before, the generic assistant answers in prose and nothing happens (the bug); after, the pause action executes and the system turn renders GovernorChatActionResult's "Paused since …" copy. Capture note: current main cannot paint this SPA at all — @loopover/engine's dist/calibration/backtest-split.js (#8097) reaches the browser graph through the chat-action registry chain and its top-level node:crypto access makes Vite's browser-external shim throw before first render, on clean main exactly as on this branch — so the harness replaces that shim with an inert stub at the network boundary; none of those node-only modules are executed by the surface being captured, and the un-stubbed behavior is identical (a blank page) before and after this change.

Viewport · Theme	Before	After
Desktop · Light
Desktop · Dark
Tablet · Light
Tablet · Dark
Mobile · Light
Mobile · Dark

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-26 05:08:36 UTC}

6 files · 1 AI reviewer · no blockers · readiness 95/100 · CI green · unknown

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This PR wires the already-built governor pause/resume chat-action stack into `conversation.tsx`'s `handleSubmit`, closing #8670 by making typed governor intents ('pause the governor', 'resume the governor') actually dispatch through the existing `runGovernorChatAction`/registry/gate pipeline instead of silently falling through to the generic streaming assistant. The new `resolveGovernorChatAction` resolver mirrors the existing portfolio-queue resolver's pure, no-best-guess pattern (word-bounded verbs, disjoint pause/resume sets, explicit unresolvable fallback for questions/ambiguous text), and `GovernorChatActionResult` is now rendered both as the pending footer and in the committed message bubble. Tests are thorough and exercise the real registry/dispatch wire end-to-end, matching the existing Ledgers-verbatim copy conventions.

Nits — 5 non-blocking

• conversation.tsx's catch block for a thrown `runGovernorChatActionImpl` (chat-conversation.test.tsx's 'surfaces a thrown dispatch' test) resets `streaming` in `finally` but never explicitly resets `governorActionPending`; worth confirming `runGovernorChatAction`'s own implementation always calls `onPending(false)` in a finally even on throw, since if it doesn't the pending footer could get stuck on a real network failure.
• fixtures.ts:20 pulls in `GovernorPauseStateResult` from `../../lib/governor` purely for the new optional field — fine, but consider whether `content` duplicating the same copy as `governorActionResult` (both carry `formatGovernorPauseChatMessage` output) is worth a comment noting it's intentional for wire-history/plain-text parity rather than accidental duplication.
• the cross-package relative import in chat-governor-resolve.ts (`../../../../packages/loopover-miner/lib/chat-governor-actions.js`) is fragile to file moves; consider re-exporting the action-name constants from the existing `@​/lib/chat-governor-actions` module already imported in conversation.tsx instead of a second deep relative path to the same underlying package.
• Add an assertion in the 'surfaces a thrown dispatch' test that the pending governor footer/status text is no longer present after the failure, to lock in that the pending flag is actually cleared on the error path.
• Consider factoring the GOVERNOR_PAUSE_CHAT_ACTION/GOVERNOR_RESUME_CHAT_ACTION re-export in chat-governor-resolve.ts through the already-imported `@​/lib/chat-governor-actions` in conversation.tsx to avoid two different import paths reaching the same constants.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#8670
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 94 registered-repo PR(s), 44 merged, 4 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	tryeverything24; Gittensor profile; 94 PR(s), 4 issue(s).
Improvement	✅ Minor	risk: clean · value: minor

Linked issue satisfaction

Addressed
The diff adds a new resolveGovernorChatAction text resolver mirroring the portfolio-queue pattern, wires it into conversation.tsx's handleSubmit to dispatch through runGovernorChatAction before falling through to streaming, renders GovernorChatActionResult in both message-bubble.tsx and the pending footer slot, and includes a substantial RTL end-to-end test that types 'pause the governor' and asse

Review context

• Author: tryeverything24
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: not available
• Official Gittensor activity: 94 PR(s), 4 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Start here: Triage stale or unlinked PRs.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

Visual preview

Route	Viewport	Before (production)	After (this PR's preview)	Diff
/	desktop	before /	after /	—
/	mobile	before / (mobile)	after / (mobile)	—

Click any thumbnail to open the full-size screenshot. Before = production · After = this PR's preview deploy.

Scroll preview

Route	Before (production)	After (this PR's preview)
/	before / (scroll)	after / (scroll)

A short scroll-through clip (desktop) — click either thumbnail to open the full animation. Evidence for scroll-linked behavior a single screenshot can't show.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.