Move ActiveAgentsPanel between AgentMetricsBar and the main agent
collection so live work is visible without scrolling past the whole
agent directory. Add a "Next heartbeat in Xs" line to each card
footer (or "Heartbeat overdue" when negative), driven by
runtimeConfig.heartbeatIntervalMs with the dashboard default fallback.
Give cards a pointer cursor + hover/focus styling so the existing
click-to-select behavior is discoverable.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Agents in `active` state without a current taskId had no SSE to attach
to, but the card still rendered "Connecting..." — implying a network
state that never resolves. Branch on taskId first: idle agents show
"Idle — no task assigned"; running-without-task shows "Starting...".
Also fix a related SSE multiplexer race: subscribers joining a channel
that had already opened never got an onOpen callback (EventSource only
fires `open` once), so they sat at isConnected=false forever whenever
another component was already streaming the same URL. Fire onOpen on a
microtask for late-joiners.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The in-review timer chip was driven by per-step instrumented duration,
which is frozen during a merge (the merge phase isn't tracked as a
workflow step). A stuck merge could read "3m" indefinitely. While
status is "merging"/"merging-pr", show live elapsed since updatedAt
(set by the merger at status flip) with a "Merging Nm" tooltip.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When the tunnel was started before parseReadiness flipped to URL-required
(or by an external invocation), the manager's status URL stays null and
buildRemoteLoginUrlForTokenType used to fall back to a constructed
http://<configured-hostname>:<port> URL — which the user saw encoded
into the QR as e.g. "fusionstudio:4040".
Add a tailscale status --json fallback that recovers the public tailnet
DNS name and reconstructs the funnel URL (https://<machine>.<tailnet>.ts.net/).
If both sources fail, surface REMOTE_URL_NOT_READY rather than emitting a
non-public URL into the auth/QR payload.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Override the stored tailscale targetPort in the start handler using the
actual port the start request landed on (req.socket.localPort), instead
of trusting a value the browser supplied. window.location.port can be
wrong behind a reverse proxy, ssh forward, or split deployment, and
silently funneling the wrong process is worse than the prior
"port-not-configured" error.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Drop the Target port field from Settings — the dashboard already knows
what port it's serving on, so use window.location.port (falling back
to 80/443 by protocol). A user-overridable port adds no value and a
stale stored value silently misroutes the funnel.
* Restructure the running-tunnel scan block to show three rows: the
raw tailnet URL, the remote access code (rt= token, separate), and
the QR image. Previously the auth URL was rendered as one long string
that read as "the bare hostname" when truncated.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Preserve and emphasize active agent card highlights across hover and modal states, with dedicated CSS updates and regression tests
- Expand dashboard and TUI activity surfaces for grouped tool-call and live agent status/log visibility, including Quick Chat and chat view refinements
- Add route and runtime updates for agent/settings flows plus self-healing coverage for merge recovery behavior
- Include accompanying changesets and test stabilizations for mission activity assertions and active highlight hooks
Fusion-Task-Id: FN-2932
- Inline grouped tool-call rendering in chat surfaces and update summary logic for clearer aggregation
- Refine chat, quick FAB, settings, and agents UI styles/behavior with matching test coverage updates
- Extend dashboard TUI state/controller flow with status and shortcut improvements
- Update engine and dashboard route handling related to agent runs, settings memory sync, and remote access adapters
Fusion-Task-Id: FN-2926
* Replace placeholder /remote/qr SVG (URL drawn as text) with real QR
rendered via the qrcode package; add format=terminal returning ASCII
QR for the TUI.
* Resolve the public tailscale funnel URL from captured CLI output
instead of constructing http://<hostname>:<port> from a configured
hostname label — that label was never used by `tailscale funnel` and
produced a non-public URL in the auth/QR link.
* Drop hostname requirement from engine + UI; only target port matters.
* Tighten tailscale parseReadiness to require a URL on the matched line
so the tunnel manager doesn't lock in `running` before the URL line.
* TUI: poll remote status, show ● tunnel indicator + URL in MainHeader,
bind Ctrl+Q to a global QR overlay (terminal ASCII), and switch the
in-Settings K shortcut to render the same ASCII QR.
* Auto-poll remote status in the dashboard while in `starting`/`stopping`
so the UI flips to running without reopening the modal.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Empty hostname triggered provider_not_configured even though the field
was labeled "(optional)". Drop the misleading label and default to the
browser's current hostname so the start action succeeds out of the box.
Also drop the legacy MediaQueryList.addListener fallback to clear the
deprecation warning.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add `triage` to VALID_TRANSITIONS["in-review"] so the dashboard's
`Request AI Revision` and `Rebuild Spec` actions work for in-review
tasks. moveTask now applies the same full reset on in-review → triage
as on in-review → todo (clears branch/baseBranch/baseCommitSha/summary/
recovery metadata and workflowStepResults) so the respec'd task starts
from scratch. The in-review task card's Move menu also gains Planning
as a destination.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When a worktree was rebased onto newer main, baseCommitSha remained an
ancestor of HEAD but the range baseCommitSha..HEAD then swept in upstream
main commits as if they were task changes (FN-2840 showed 33 files for a
4-file task). The previous display-recovery only fired when baseCommitSha
was no longer an ancestor, missing this case.
When enableDisplayRecovery is on and baseBranch is missing, also compute
merge-base(HEAD, main) and prefer it when it's a descendant of
baseCommitSha (tighter fork point). When merge-base is not a descendant
(FN-2855: baseCommitSha is on a deleted feature branch), keep
baseCommitSha so we don't widen to unrelated upstream files.
Also stabilizes the card-vs-Changes-tab divergence: resolution no longer
depends on whether a rebase happened between polls.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two fixes for the in-review/in-progress "files changed" count on task
cards and the Changes tab:
- Drop untracked files from /tasks/:id/diff and /tasks/:id/file-diffs.
At review time these are almost always build artifacts/cache/logs not
in .gitignore, not real task changes, and they inflated the count.
- Add display-only `enableDisplayRecovery` option to resolveDiffBase.
When a worktree was rebased onto origin/main after baseCommitSha was
recorded and baseBranch was not set, the prior code fell through to
HEAD~1 — undercounting to just the last commit's files (e.g. FN-2957
showed 2 in review, 6 after merge). Recovery now tries
merge-base(HEAD, main) / origin/main before HEAD~1.
Routes opt into recovery; the merger's mirrored copy
(resolveTaskDiffBaseRef) is intentionally untouched so merge-time
scope warnings still evaluate the strict task base.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Update dashboard mocks to push assistant text into session.state.messages
(matching real session.prompt() void return) and add a regression test for
array-shaped content. Add core unit tests for extractDreamProcessorResult
to lock in undefined/null safety.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
session.prompt() resolves to void; the reply lives in session.state.messages.
The dream route was awaiting its return value, so parsing crashed with
"Cannot read properties of undefined (reading 'match')" once the dream
completed. Extract the assistant text post-prompt and fail soft on
non-string input in extractDreamProcessorResult.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two pre-existing local modifications that had been sitting in the
working tree across this session, neither produced by the recovery /
fallback work in adjacent commits. Committing them so the tree is clean.
- packages/dashboard/app/components/SettingsModal.css
Removes one trailing blank line at end-of-file. No semantic change.
- packages/dashboard/app/components/__tests__/ListView.test.tsx
Drops one test case ("shows NodeHealthDot for selected bulk node
override", 15 lines) from the "ListView - Bulk Selection" describe
block. The test asserted that selecting a bulk node override
rendered a `.status-dot--online` indicator in the DOM. Sibling tests
(option labels, batchUpdateTaskModels invocation) are preserved.
Likely debris from the FN-2729 NodeHealthDot work that didn't survive
one of the in-flight squash merges; the underlying behavior is still
exercised by the kept tests.
Verified: pnpm --filter @fusion/dashboard test ListView passes
(125/125), so the removed assertion isn't load-bearing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wires the AI commit-body generator (introduced in 60217ce19) to the
existing dedicated title-summarization model lane, and updates the
settings UI so users understand the model is now used for two
short-summary jobs instead of just one. Also de-duplicates the
project-scope settings UI which was exposing the same model setting in
two places.
Engine (merger.ts):
- aiGenerateCommitBody now prefers settings.titleSummarizerProvider /
titleSummarizerModelId over the merger's default model. The
summarization lane is the right tier for this work — small, fast,
cheap. Falls back to the default merger model when the summarization
lane is unset.
Dashboard (SettingsModal.tsx):
- MODEL_LANES summarization lane label updated:
"Title Summarization Model" → "Title and Git Commit Message
Summarization Model". Helper text updated to mention the dual purpose
(auto-generated task titles + fallback merge commit message bodies).
This change flows through automatically to BOTH the global model
lanes view and any project-scope rendering of MODEL_LANES.
- Removed the duplicate summarization picker from the project-scope
Model Lanes section: previously it appeared once under Model Lanes
AND once under "AI Summarization". Now lives only in the dedicated
picker so users have a single source of truth in project scope.
- The dedicated picker section heading + description rewritten:
"AI Summarization" → "AI Title and Git Commit Message Summarization",
with a paragraph explaining both jobs the model performs.
- Inner dropdown label updated for consistency.
Tests + checks: engine 2887/2887 pass, dashboard typecheck clean,
workspace lint clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Resolve merge conflicts in InlineCreateCard and QuickEntryBox using the existing node picker UX
- Replace ad-hoc node status rendering with NodeHealthDot in task creation surfaces
- Keep advanced quick-create controls and portal-based dropdown behavior intact
- Preserve related dashboard styling and tests included in the staged squash
Fusion-Task-Id: FN-2729
Self-review of the recovery cascade surfaced three issues; this commit
addresses all of them.
1. AI didn't actually receive the safety constraint under Layer 3.
The previous commit logged the safety preamble to the task log via
`store.logEntry`, but the merge agent doesn't read task log entries as
prompt context — so the AI was running blind. The "no silent
re-introduction of main's deletions" guarantee was therefore relying
*entirely* on the deterministic verification gate (test + build),
which is correct as a backstop but doesn't help the AI produce a
correct first attempt.
Fixed by threading `preMergeRebaseFallthrough` through
`MergeAttemptParams` → `executeMergeAttempt` → `runAiAgentForCommit` →
`MergePromptParams` → `buildMergePrompt`, where it now injects an
explicit "⚠️ Pre-merge rebase recovery exhausted" preamble at the top
of the user prompt with three concrete rules:
- Prefer main's deletion when branch re-adds removed lines
- Prefer main's version on ambiguous hunks
- Call `fn_report_build_failure` rather than commit a regression
Also includes the original rebase failure message (truncated) so the
AI has diagnostic context.
The truncated-context retry path also forwards the preamble — it's the
safety constraint, not bulk context, so we keep it even when stripping
diff stat / commit log to fit the window.
2. Layer 2's branch-restore could fail with "uncommitted changes".
When a cherry-pick midway through Layer 2's replay fails, the worktree
is in a half-applied state with conflicts in the index. The previous
restore did `git checkout <branch>` (no -f) followed by
`git reset --hard <originalSha>`. The plain checkout would refuse with
"would overwrite local changes" if there were unmerged paths,
preventing the reset from running and leaving the branch at the
half-replayed tip.
Fixed by reordering: hard-reset to the captured original SHA first
(this clears index/working tree of any cherry-pick state), then
`git checkout -f <branch>` to ensure HEAD points at the named branch,
then a final hard-reset to the original SHA as belt-and-suspenders.
Worst case the worktree is at the original branch tip — never worse
than where Layer 2 started.
3. Pre-existing unrelated lint error blocking workspace lint.
`packages/dashboard/src/server.ts` had an unused `resolve` import from
`node:path` left behind by a recent refactor that extracted
`PACKAGE_VERSION` into its own file. The user explicitly asked to
clean it up so workspace lint passes. One-line drop.
Tests + checks:
- Engine: 2886 / 2886 pass (added safety preamble didn't break any
existing prompt-content assertions)
- Core: 3120 / 3120 pass
- Workspace lint: clean
- Engine typecheck: clean
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
