Lab #3108: Resume signalled workflow runs automatically under swamp serve

Problem

Split out of swamp-club#3094, which now covers only delivering a signal through swamp serve. After it, a signal can arrive over WebSocket or HTTP, but the run still sits suspended until someone resumes it by hand, also when serve started the run.

Serve has no way to learn that a signalled run can continue:

A signal creates an outcome record and never changes the run record (swamp-club#3093). WorkflowRun.isAwaitingResume() stays false while any signal wait is on the record, settled or not, so the derived awaitingResume is unset.
autoResumeAfterApproval in src/serve/resume_launcher.ts is called only from the approve handler.
autoResumeParentAfterChild returns false outright when the parent has a signal wait.
Nothing in serve sweeps for suspended runs. A lost auto-resume launch is logged and audited and never retried.

Two serve instances on one datastore could also both launch the same resume: the active-run registry is per process, and serve resumes with unclaimedRuns instead of the lock-backed run claim.

Proposal

Decide resumability from the run record together with the outcome records: suspended, no gate undecided, no nested run waited on, and an outcome for every wait. isAwaitingResume(store, run) in src/libswamp/workflows/signal.ts already computes this for the signal result and can move into the domain.
Reuse the workflow's existing autoResume setting; no second default for signals.
Validation requires a workflow that contains a wait and declares inputs to set autoResume explicitly, so no signal-driven run silently waits for a resume nobody knows to send.
A continuation claim, created with putIfAbsent per run and suspension, so that two serve instances never launch the same resume. A claim held by an instance with no live heartbeat is released and competed for again.
The instance that accepts a signal attempts the continuation. It may not have the run record (runRecordAvailable: false), so the sweep is the path that always works.
A continuation sweep at boot and on an interval: find suspended runs whose waits are all settled and whose policy allows, and launch them through the claim. Capacity, shutdown and authorization refusals leave the run suspended with a visible reason, and the next sweep retries.
The resume runs as the run's initiator, re-authorized when it launches. The signal's submitter is an audit actor only.
A parent suspended on a child that waited for a signal is auto-resumed when the child ends, as it is for a child that waited on a gate.
On resume takeover, pull the run record under the claim when the datastore is remote, and have serve take the lock-backed claim for this path.

Open decisions

Where continuation claims live. Wait records are in the datastore's own control-plane store. Serve's other control-plane records, heartbeats included, fall back to a directory local to each repository on a filesystem datastore. Dead-holder detection needs claims and heartbeats visible to every instance.
How a resume is authorized as the initiator. initiatedBy on the run is a plain string, and decideSubjectAccess refuses a subject with no token binding. The cron trigger authorizer is the nearest precedent.
Whether the sweep acts on a custom datastore, where this host's copy of a run record can be missing or behind. The wait-record cleanup sweep refuses to act there for that reason.
Whether the sweep also resumes approved runs whose auto-resume launch was lost. That is a fix, but it changes when such runs execute.

Acceptance criteria

A signal that settles a run's last wait resumes the run without a manual resume when autoResume allows, and exactly once with two serve instances on the same datastore.
A resume launch lost to a crash or a capacity refusal is retried by the sweep.
A continuation claim left by a dead instance does not strand the run.
A run whose initiator lost the authority to run the workflow is not resumed, and its receipt stays readable.

Tests

Integration tests with ordered creates, never sleeps: two instances competing for one continuation, a claim left by a dead instance, a launch refused for capacity and retried.
integration/workflow_run_claim_rules_test.ts updated for the resume change.

Out of scope

Settling overdue waits (swamp-club#3109), the client and dashboard display (swamp-club#3110), leases with fencing for executors, continuing one branch while another still waits.

References

swamp-club#3094 (delivery), swamp-club#3093 (outcome records), the Wait for Signal section of design/primitives/workflows.md.

Session-start 'swamp model search --json' in the AGENTS.md template is unfulfillable - wire it to a hook or drop it

Flaky property test: analyzeExpression dot/bracket agreement fails when the generated model name is a CEL keyword

Accept usernames for user subjects in grant files

Docs: enable-managed-config should say to re-run config migrate when its push fails

Docs: grant file reference shows a stale format and omits resources and subjects

Flaky property test: analyzeExpression hyphenated dot access fails when modelName generates a CEL keyword ("in")

extension quality and push crash with IsADirectory when additionalFiles lists a directory

extension push: a required global argument renders with a stray "" ("text": "string""")

Docs: extension publishing reference is stale on bare specifiers, quality output and fmt config (swamp-club#3022)

swamp-extensions: replace bare import specifiers with npm:/jsr: specifiers in first-party extensions

extension push: 'Requested visibility' line prints to stdout without the logger prefix, twice per run

A nested structural swamp that outlives the run that started it keeps skipping that run's lock while the holder writes

Show signal waits to remote clients and the dashboard

Settle expired signal waits from swamp serve

Resume signalled workflow runs automatically under swamp serve

extension push/quality: structured acceptances for agents, a clearer title than For next time, and an optional reason

extension push: the undecided private-entitlement verdict still says the registry may start a trial at publish

seed:collective cannot produce a superseded personal trial or a returned trial, so the UAT cannot cover them

Replayable state fragments: Git-reviewable desired state for resources changed by model methods

extension push: registry metadata and content hash label files relative to the repo dir, so a push with --extensions-dir records ../ paths

registry: store and show declared acceptances sent in contentMetadata.acceptances at confirm (swamp-club#3021)

extension push --dry-run: authentication verdict prints 'Signed in as .' when the credential is an API key with no username

serve --remote-only: step output from workers is never persisted to the run log (history logs always empty)

serve HA: worker enrollment on a replica that lacks the enrollment token's definition may create a second definition (unverified)

serve HA: pull a token's definition on an auth miss so a peer accepts a new token at once

extension push: version-drift check reports 'no previously published version found' when the registry call failed

Search and list output follows directory iteration order in the definition, workflow, vault and output repositories

extension push: accept attestation-backed adversarial-review evidence (lookup by extension and content hash), generated provenance, and carry the evidence to the registry at confirm (design: Lab #3023)

Workflow load errors print the raw Zod issue dump instead of a readable message

Decide a stricter naming rule for workflow step and job names

serve: run scheduled workflows concurrently (per-workflow serialization), and report queue delay

Show expired approval gates as expired, with a Cancel action in the dashboard

digitalocean codegen leaves an orphaned security_secret.ts model file after its endpoint left the spec

extension promote accepts a yanked version and reports it promoted

Docs: doctor install reports a stale or failing autoupdate scheduler

extension push output: double-quoted adjacent interpolations, repo-relative paths, and --json emits three JSON documents

auto-resolve: the Installing line prints the extension's entire multi-line description

s3-datastore/serve: readResource without a version returns an older version; the latest marker is rewritten backwards

Publish agent-facing projections of the manual: /llms.txt, .md pages, and llms-full.txt

Let agent-runner drive more agent CLIs (Kilo Code and others)

s3-datastore: every fast-path miss re-downloads every _index shard, so a busy serve pulls the whole index every poll (~230 GB/day S3 egress)

Feedback: Swamp as an approval-gated control plane for a small fleet

extension safety analyzer's Deno.Command( warning is a plain substring match

Tell timeouts apart from cancels in method-run records, and review the hidden 30 s fallback timer for step-called model methods

Run from a git worktree: definitions from the worktree, state from the shared repo

S3/GCS datastore: a push that fails after uploading drops its recorded deletes, so the retry brings deleted data back

Tests: bring the in-memory remote's default semantics up to @swamp/s3-datastore and @swamp/gcs-datastore 2026.10.01.1

Tell the pushing client when its extension contentMetadata fails validation

Extension datastore: query still returns an item after its delete is pulled

Cancel, reject and supersede of a parent run should settle its suspended nested child runs

Relationships

#3108 Resume signalled workflow runs automatically under swamp serve