Same class of failure as the kb-db-test-* leak: vitest's forks pool
SIGTERMs a fork on test timeout and skips the in-test `finally { rmSync }`,
leaking `fusion-test-ref-project-*`, `fusion-test-ref-concurrent-project-*`,
and `fusion-test-ref-advance-*` dirs that scripts/check-test-isolation.mjs
flags during deterministic merge verification.
Track every minted dir in a per-file set and sweep it from
SIGTERM/SIGINT/SIGHUP/beforeExit/exit handlers (signals re-raised after
cleanup) plus an `afterAll` for the happy path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Address the code-review finding on the prior commit: a natural premise-stale
summary like "PREMISE STALE: the task has no remaining work — implementation
is already done on HEAD" matches /\b(incomplete|not implemented|not done|
not finished)\b/i with 'the task' inside the 40-char first-person window,
refusing fn_task_done with summary-claims-incomplete and deadlocking the
escape hatch.
When summary starts (case-insensitive) with PREMISE STALE:, skip the
dissent-pattern and scoped-incomplete summary checks. Pending-code-review
and bulk-step-completion guards still apply unchanged.
Add executor-task-done-premise-stale.test.ts covering: the deadlock case
now passes; dissent phrasing in a sentinel summary is allowed; case-
insensitive sentinel; sentinel must be at the start (mid-summary doesn't
bypass); REVISE verdict still blocks even with the sentinel.
Root-cause prevention for the FN-5521 failure class where PROMPT.md describes
work that is already done on HEAD and the executor still marches through
plan/review/test/doc, then merger burns hours retrying against unrelated
flakes in @fusion/engine reliability-interactions.
- Executor system prompt: add a Preflight escape hatch. When Step 0
reproduces and finds HEAD already matches the desired state, the agent
marks Step 0 done, marks remaining steps skipped, and calls fn_task_done
with a `PREMISE STALE:` summary. Reuses the existing
evaluateTaskDoneRefusal/skipped semantics and the merger's
empty-own-diff fast-path — no new tools or refusal classes needed.
- packages/engine/vitest.config.ts: split into two projects. engine-default
retains full parallelism; engine-reliability scopes
src/__tests__/reliability-interactions/** to
poolOptions.threads.singleThread so event-ordering assertions stop
flaking under workspace-concurrent merge-gate load (the
`expected 24 to be less than 19` rowid interleaving in
merge-reuse-task-worktree).
Two test-infrastructure fixes from agent #1's flake-stabilization pass.
Skipped its DESC-order assertion rewrite (semantically equivalent no-op
that conflicted with the prior boilerplate consolidation in 848a226ca)
and its vitest.config additions (already in flight on main).
1. vitest-setup.ts: completedSubprocessFailures was a plain string[]. When
a 30s subprocess-guard timer fired during a *later* test's execution
window (because the owning test ran for e.g. 40s under its 60s
timeout budget), the failure surfaced in the innocent successor test's
afterEach. Typed the array as { ownerTestName, message }[] and filter
on the current test name; orphaned entries are dropped silently.
2. worktree-contamination-attribution.real-git.test.ts: afterEach rm
occasionally hits ENOTEMPTY on macOS when a git rebase internal dir
isn't fully flushed. Added maxRetries: 3, retryDelay: 100.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Code-review fixes for 036387610 / d02cd38d7:
1. `deriveScopedPnpmTestCommand` now passes each `--filter` argument through
`quoteArg`. Package names come from workspace `package.json` files which
are not strictly trusted input — a metacharacter in a name would have
leaked into the shell command.
2. `getBranchChangedFiles` now quotes both git refs in the `<base>...<head>`
range. Branch names can legally contain `/` and other characters; this
is defense-in-depth consistent with the rest of merger.ts.
3. Out-of-scope detection is now package-aware via a new
`packageNamesForFiles` helper. A failure in `__tests__/foo.test.ts` is
correctly treated as in-scope when the branch touched `src/foo.ts` in
the same package, whereas the previous filename-prefix heuristic missed
that case entirely. The dead `bf.startsWith(ff/)` clause is removed.
Falls back to the directory-prefix heuristic when pnpm-workspace.yaml
is unavailable.
195 tests pass across merger-verification.test.ts and the reused
merge-reuse-task-worktree.test.ts.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Each of the 18 tests in this suite was ~50 lines of identical setup: rename
main→master, mark steps done, create the branch with one content commit,
create the task worktree, set task.worktree, enqueue the merge queue. The
test bodies were dominated by ceremony and the unique assertions were buried.
Consolidated into a single `setupReuseHandoff` helper that takes per-test
overrides (skipWorktreeAdd, worktreeOverride, skipEnqueue, emptyOwnDiff,
extraSettings). Added a top-level beforeEach to clear shared mock/registry
state that several tests forgot to reset.
No behavioral change. All 21 tests pass on two consecutive runs in ~20s.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Follow-up to a3ec2e55c addressing code-review feedback:
- Restrict the existence check to local heads via
`git show-ref --verify --quiet refs/heads/<name>` so a stray tag or
remote ref with the same name cannot satisfy the check and lead the
subsequent `git checkout` to a detached HEAD on the wrong object.
- Use `git branch -M` (force) instead of `-m` for the rename path so
case-only renames succeed on case-insensitive filesystems (macOS,
default Windows).
- Document the FN-5456 invariant on the function's doc comment.
- Add a real-git regression suite covering the three paths the fix
governs: no creation when expected ref is missing, switch-only when
it exists, and rejection of same-named tags.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`attemptBranchAutocorrect` was the only branch-creation site in the
engine that ran `git checkout -B <expected>` without a start point.
When the worktree's HEAD was at a previous occupant's tip, the new
label silently captured that commit — the "branch: Created from HEAD"
contamination pattern that the cross-contamination guard then refuses
to auto-resolve (see FN-5456: orphan FN-5477 commit 268574b9a stranded
on fusion/fn-5456).
Replace the unsafe fallback with verify-then-`git checkout`: only switch
to an existing expected ref; return `failed` when it does not exist so
upstream recovery — which knows the proper base SHA — can re-anchor via
`prepareForTask` / `reanchorBranchToBase`.
Tests updated for the new command sequence; new case covers the
"expected ref missing → fail without creating from HEAD" guarantee.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When the in-merge fix agent makes no changes AND all failing test files are
outside the branch's diff, the merger now throws OutOfScopeVerificationError
and marks the task status: "failed" with a clear error message:
"Merge verification failed in files outside branch scope — likely
pre-existing flake on main. Fix the base-branch test breakage
separately and retry."
This prevents the task from entering the completion-handoff-limbo recovery
cycle (which would retry the merge endlessly) when the verification failure
is caused by pre-existing flakiness in an unrelated package (e.g. engine
reliability-interaction tests failing while only dashboard was changed).
Failing file paths are parsed from vitest/jest output (FAIL lines and ❯
summary lines). If parsing yields no file list, the existing retry behavior
is preserved. The OutOfScopeVerificationError propagates through the catch
block so it does not count toward completionHandoffLimboRecoveryCount.
New exports: OutOfScopeVerificationError, parseFailingFilesFromOutput,
getBranchChangedFiles.
Tests added: parseFailingFilesFromOutput (4), getBranchChangedFiles (3),
OutOfScopeVerificationError constructor (1). All 58 merger-verification
tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When a pnpm workspace is detected and git context is available,
inferDefaultTestCommand now derives the set of packages touched by the
branch diff (git diff --name-only <base>...HEAD) and emits:
pnpm --filter "<pkg>...^" test
instead of the broad `pnpm test`. The `...^` suffix includes dependents
so packages that import the changed one are also exercised. Falls back
to unscoped `pnpm test` when git context is missing, the workspace has
no package roots, or all changed files are at the root (e.g. config).
New exports: parsePnpmWorkspaceGlobs, resolveWorkspacePackageRoots,
mapChangedFilesToPackageNames, deriveScopedPnpmTestCommand.
testSource is now "inferred-scoped" for the scoped path.
Tests added: parsePnpmWorkspaceGlobs (7), resolveWorkspacePackageRoots
(4), mapChangedFilesToPackageNames (4), inferDefaultTestCommand scoping
(6). All 466 merger test files pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Switch commits Fusion produces (both executor step commits and merger squash
commits) from setting `--author="Fusion <noreply@runfusion.ai>"` to appending
`-m "Co-authored-by: Fusion <noreply@runfusion.ai>"`. The user's configured
git identity now stays as the primary author/committer, and Fusion is recorded
as a co-author (recognized by GitHub for shared attribution). The
`commitAuthorEnabled` toggle and `commitAuthorName`/`commitAuthorEmail`
settings keep their existing keys; the dashboard settings UI relabels them
from "Author" to "Co-author" to match.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two engine merger tests created mkdtempSync workspaces directly in tmpdir()
under the tracked `fusion-test-` prefix; under full-suite concurrent load
the post-run check-test-isolation flagged them as leaks. Route both
(`merger-no-op-fix-finalize.test.ts`, `merger-verification-fix-already-on-main.test.ts`)
through FUSION_TEST_WORKER_ROOT like sibling merger tests so they nest
inside the already-tracked worker root.
Bump engine vitest subprocess guard from 60s to 120s and testTimeout to
30s — plain git commands (branch -d, worktree remove) queued behind
system contention during `pnpm -r --workspace-concurrency=2` runs were
timing out. The guard only fires on hangs, so healthy tests pay nothing.
Also bundles in-progress dashboard mobile-breakpoint regex/CSS test
updates and docs index additions.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Code-review follow-ups on 24686cade:
- Only `git reset --hard HEAD` / `git clean -fd` after the stash is confirmed (non-empty SHA + store succeeded). Previously an empty `git stash create` result (no throw) would still trigger the reset and wipe the user's dirty edits before the refusal was raised.
- Quote the stash label with a local `quoteAutostashArg` helper to match `merger.ts`'s `quoteArg` pattern. Task IDs are alphanumeric today, but defense-in-depth keeps `git stash store -m` safe against future label content.
- Add `merge:reuse-handoff-autostash` to the typed audit unions in `run-audit.ts` and `emitReuseHandoffAuditEvent`, so subscribers see it as part of the closed set instead of relying on the `as any` escape hatch.
- Best-effort `git reset` (unstage) in the autostash-failure path so the worktree isn't left half-staged when we refuse.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Stop losing uncommitted dev edits during task merges.
- `acquireReuseHandoff` no longer throws MergeHandoffRefusedError("working-tree-dirty") on a dirty reused worktree (FN-5138). It autostashes via `git add -A` + `git stash create` + `git stash store`, emits a `merge:reuse-handoff-autostash` audit event with the stash SHA and a recover command, and lets the merge proceed.
- `stashUnrelatedRootDirChanges` no longer silently proceeds when stash creation fails on a dirty tree. It throws a new `AutostashCreationFailedError`; the merger catches it and surfaces a clear "your edits are intact" message before any destructive op runs.
- New failure reason `dirty-worktree-autostash-failed` distinguishes stash failure from the old refusal.
- Tests in `merger-integration-worktree`, `merger-cwd-fallback-removed`, and `reliability-interactions/{integration-worktree-state,merge-reuse-task-worktree,cwd-integration-fallback-removed}` updated to the new contract; the FN-5348 "no cwd fallback" invariant is preserved.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a 255-line fast-path regression harness (`self-healing-fn-5488-fast-path-regressions.test.ts`) covering unbacked merge fast-paths, retry-exhausted fanout, grace-window boundaries, and overlap preservation invariants. Also includes a small cleanup to `MobileNavBar.test.tsx`.
Fusion-Task-Id: FN-5498
Fusion-Task-Lineage: f39ca8a1-223b-4b43-a844-6cb17d19e099
Fusion-Task-Id: FN-5498
Added soft-delete blocker recovery logic to the scheduler and self-healing systems, enabling reconciliation of stale blocker reasons when tasks are archived or restored. The changes include corresponding tests for the completion guard behavior, scheduler recovery paths, and self-healing integration,
Fusion-Task-Id: FN-5496
Surface the live mode / modalOpen / keyboardOpen / footerVisible / view
values that MobileNavBar uses for its early-return so the ?vpdebug overlay
can show which one is hiding the bar on Android. Also dumps the
.project-content className so we can correlate with `--with-mobile-nav`.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds an oscillation backstop and guard mechanism for completed tasks to the self-healing system, including a dedicated 145-line test suite for completed-task oscillation behavior and a small update to the existing self-healing module.
Fusion-Task-Id: FN-5448
The merge delivers the FN-5483 identity-guard bypass for merger-driven commits on detached HEAD, plus Steps 2 and 6 of FN-5479 which gate the limbo counter by enqueue acceptance and document the associated invariant. It also restores the dashboard's PWA and theme-boot contract in index.html. New reg
Fusion-Task-Id: FN-5479
WorktreePool.prepareForTask now rejects empty/"HEAD" base values and verifies
that the worktree's HEAD actually landed at the resolved base SHA after
`git checkout --detach`. This closes the FN-5432 / FN-5255 contamination
pattern where a recycled worktree branched from a stale HEAD (reflog:
"branch: Created from HEAD") and pinned the new task's tip to the previous
occupant's commit.
SelfHealingManager.tryReanchorForeignOnlyContamination is invoked from both
PR-conflict and self-owned-branch-conflict catch sites before the
unrecoverable-pause path. When the conflicted branch carries only foreign
commits (no own work), the branch is reset to base via the existing
recoverForeignOnlyContamination flow instead of being escalated for human
adjudication.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Completes FN-5481 by unskipping soft-delete matrix backstop tests, adding substantial new coverage to the soft-delete end-to-end test suite (290 lines added/modified) and updating the verification matrix documentation to reflect the newly enabled test scenarios.
Fusion-Task-Id: FN-5481
- Add core settings schema/types support for testMode with model-resolution override handling
- Enforce engine session lane overrides in test mode with targeted helper coverage
- Add dashboard settings toggle plus persistent test-mode banner and related component tests
- Update settings documentation and parity/roundtrip tests for the new test mode behavior
Restructured AGENTS.md from a 588-line catch-all into a lean reference of essential rules by offloading deep guidance into five new/expanded doc files: agents.md, architecture.md, dashboard-guide.md, settings-reference.md, and testing.md. Also restored a missing FN-5345 lifecycle invariant in self-h
Fusion-Task-Id: FN-5339
Fusion-Task-Lineage: e1a2dc4f-ae50-46f4-acd3-8f2c3673feb4
Contamination on fusion/<id> branches (FN-5233 was the recent example:
two untrailered feat(FN-5353): commits sitting on fusion/fn-5233) used
to be invisible until merge time, days after it happened. The executor
already runs assertCleanBranchAtBase at worktree acquisition and reclaim
— the gap was the active session window itself.
Add reportBranchAttribution(repoDir, branch, baseSha, taskId) which
walks base..branch and bins every commit into ownTrailed (healthy),
ownUntrailed (subject tag but commit-msg hook didn't fire), foreign
(different FN-id), or unattributed (no subject pattern, no trailer —
typically a hand-merge or plumbing commit). Wire it into the executor
right after captureModifiedFiles in the post-session path: when any
anomaly bucket is non-empty, emit a structured branch:attribution-
anomaly audit event and a task log entry. The audit itself is wrapped
in a try/catch so a probe failure never destabilizes a completing
session. New branch:attribution-anomaly and branch:auto-reattach-
authoritative GitMutationType variants accept the structured metadata
(the latter for the handoff re-attach added earlier this session).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
commitOrAmendMergeWithFixes used to swallow every unexpected throw as
`reason: "unknown-phantom"` and the two callers re-threw a bare
"verification fix finalize failed (unknown phantom)" with no operator-
actionable signal. FN-5422-class wedges (preAttemptHeadSha == currentHead
but branchTip ahead with task-trailered commits) couldn't be diagnosed
without re-running.
The catch now records the original error and probes whether the branch
ref itself is authoritative for the task (tip carries Fusion-Task-Id
trailer, base..branch is foreign-contamination-free). When it is — i.e.
the work isn't lost, the integration worktree just didn't advance — the
catch resets rootDir to preAttemptHeadSha so the next merge attempt
starts from a known baseline instead of inheriting partial squash state,
and returns `reason: "branch-ref-ahead-reset"`. The two callers fold
`originalError=` and `branchAuthority=` into the thrown message so the
real cause (diff-volume gate, file-scope, transient git, etc.) is
visible in the run log.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The merge handoff refused with head-branch-mismatch whenever the reused
worktree's HEAD wasn't on fusion/<id> (detached, recycled to main, or on
a sibling branch), even when the branch ref itself still held a clean,
task-attributed lineage. That wedged FN-5339-class tasks in review for
no good reason.
Add isBranchAuthoritativeForTask in branch-conflicts.ts (branch ref
exists, tip carries Fusion-Task-Id trailer, base..branch is foreign-
contamination-free) and use it in acquireReuseHandoff: when HEAD drifts
but the branch ref is authoritative, run a plain `git checkout <branch>`
inside the already-asserted-clean worktree, re-read HEAD, and emit a
branch:auto-reattach-authoritative audit. Refusal still fires unchanged
when the branch ref is missing, missing the trailer, or contaminated,
so FN-5363 strict-lease and foreign-commit guards remain authoritative.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a manual merge blocker mode (FN-5438) that prevents automatic merging and provides a bypass mechanism to resume, wired through the merger, project engine, and task workflow API routes. Includes tests across core, engine route registration, and project engine layers, plus a changeset and documen
Fusion-Task-Id: FN-5438
Implements stash conflict detection and resolution with a new smart pull API route, resolution endpoints, and a dedicated conflict modal UI — backed by extended git audit taxonomy and route tests.
Fusion-Task-Id: FN-5358
Merges the Layer 2.5 scope-auto-widen feature (FN-5226) into the merger: a new evaluator module that automatically widens a task's declared file scope based on git attribution prior to the existing scope partition gate, wired into `merger.ts` with full audit taxonomy, persisted task metadata, and re
Fusion-Task-Id: FN-5226
Test coverage hardening for FN-5440: adds stronger interaction tests across three steps covering the cwd-integration-fallback-removed path and the merge-reuse-task-worktree early fast-path, including expanded fallback audit tripwires and hardened metadata literal guards. No production code changes.
Fusion-Task-Id: FN-5440
Removes the broad-scope detection feature end-to-end: the TaskCard chip and TaskDetailModal advisory banner are gone from the dashboard, the triage heuristic that flagged tasks as broad-scope has been deleted from the engine along with its associated run-audit events, and documentation references ha
Fusion-Task-Id: FN-5405
FN-5351 adds structured telemetry for merge audit events and integration worktree state, including typed ref-advance tracking, terminal handoff fallback audit, and integration state probes with corresponding reliability backstop tests and documentation updates.
Fusion-Task-Id: FN-5351
Implements ref-only integration branch advancement via `git update-ref` instead of checkout+commit, including a new `merger-ref-update-advance.ts` helper, a reduced and clarified `merger.ts`, comprehensive unit and real-git regression coverage, and audit event wiring.
Fusion-Task-Id: FN-5350
Removes the `cwd-main` integration fallback mode (FN-5348), eliminating the legacy shortcut path where the merger would operate directly on the project root instead of a dedicated worktree. Steps normalize the `reuse-task-worktree` integration mode as the sole path, wire stricter mode invariants in
Fusion-Task-Id: FN-5348
Fixes a stale queued-status recovery branch in self-healing and updates the corresponding test assertions to match the corrected behavior.
Fusion-Task-Id: FN-5434
Added executor logic to skip retries when a review is pending for a task, introducing a `pendingReviewBlockHelper` in the task-done path and updating the retry-gate to consult it; two new reliability-interaction test suites cover the feature behavior and composition with existing retry/backstop laye
Fusion-Task-Id: FN-5436
