* Add runner label to status output

* Add changelog entry for status runner label

* Fix status runner detection and sanitization

(cherry picked from commit 03477cc)

* Add runner label to status output

* Add changelog entry for status runner label

* Fix status runner detection and sanitization

(cherry picked from commit 03477cc)

…e trap-class

Three confirmed instances from #325 phase-2 rebase:
- e515ea1 (gateway live test hardening; base has richer probes via f07b00d+a53fea3905+5f702b464b)
- aa1908b (docker live backend; same parallel-evolution pattern)
- 7ee46a3 (PR openclaw#70595 already in base CHANGELOG; base evolved naming + lookup table)

Discovery channels: git cherry (insufficient), CHANGELOG-byte-grep (high precision), conflict-content rubric.

Tracking: #331
Eats dogfood for: openclaw-bootstrap#702

…y primitives)

Wires the §1 entry-point it.todo placeholder
  swim-runner.test.ts L80: "rebase bot classifies synthetic squash-rebased
  commit as DROP (not PICK)"
onto runnable ground by composing the two §1 discovery primitives
(changelog-grep + cherry-pick-provenance) into a single DROP/PICK/REVIEW
verdict with channel attribution and structured evidence for downstream
span emission.

API:
- classifyRebasePick({subject, commitBody, baseChangelog, isAncestorOf})
  returns {verdict: DROP|PICK|REVIEW, channel, evidence,
           needsConflictContentInspection?}

Channel ordering (memo-aligned, deterministic):
  1. cherry-pick-provenance (highest precision: exact source SHA,
     ancestor-of-base check unambiguous)
  2. changelog-grep:pr (high precision: PR tokens rarely collide)
  3. changelog-grep:subject (medium precision: substring, can collide)

Load-bearing invariant: when both positive-signal channels miss, the
verdict is REVIEW with needsConflictContentInspection=true \u2014 NOT a
silent default to PICK. The memo's third channel (conflict-content
classification rubric) needs file-path + diff inspection, not pure
string\u2192struct, so it's NOT covered here; classifier surfaces the gap
explicitly so the caller can hand off rather than misclassify the
test-harness divergence cases.

Evidence completeness: even when one channel is load-bearing for the
verdict, ALL channels' evidence is recorded in the result so the
downstream span emission can carry full audit data. Non-ancestor
cherry-pick footers go to a separate
slot for cross-branch provenance audit.

Boundary: pure function. Caller drives:
  - subject from `git log -1 --format='%s' <pick>`
  - body from `git show -s --format=%B <pick>`
  - base CHANGELOG from `git show <base>:CHANGELOG.md`
  - isAncestorOf wrapping `git merge-base --is-ancestor`

Coverage (10 tests):
- Instance 1 (openclaw#70595): DROP via changelog-grep:pr; no
  needsConflictContentInspection on DROP
- Instance 3 (cherry-pick): DROP via cherry-pick-provenance;
  precedence over CHANGELOG when both fire; non-ancestor footer
  doesn't trigger DROP via that channel
- Instance 2 (test-harness): REVIEW with
  needsConflictContentInspection=true; classifier never defaults to
  PICK on absence-of-signal
- subject-channel: DROP via changelog-grep:subject
- evidence completeness: both channels' evidence recorded when both
  fire; non-ancestor footers separated from load-bearing slot

Result: 10/10 passing this file. swim-37 vitest project: 79 passed |
18 todo (6 files). pnpm lint:core: 0/0.

Closes part of #324; third of three §1 todos wired to runnable ground.

…el (#408)

* feat(swim-37): #324 §1 trap-class CHANGELOG-byte-grep discovery channel

Wires swim-37 trap-class §1 (parallel-evolution / cherry-false-negative)
discovery channel onto runnable ground. Pure function, byte-pinned to
studies/swim-37/traps/parallel-evolution-class.md §2 byte-walk against
v2026.4.24 CHANGELOG L164 (instance `7ee46a3ab9 fix: Add runner label
to /status (openclaw#70595)`).

API:
- grepChangelog(subject, content) — literal grep -F semantics, returns
  ALL hits (memo flags subject-line collisions as known false-positive
  mode; channel returns all so caller can decide)
- extractPrNumberToken(subject) — pulls trailing `(#N)` PR-token, last
  match wins (commits sometimes cite earlier PRs in body but subject's
  trailing PR is the landing PR)
- discoverChangelogHit(subject, content) — composite: PR-token first,
  fall through to full-subject; needle field tells caller which channel
  found the hit (or which was attempted last on miss)

Wires trap §1 it.todo:
  swim-runner.test.ts L81: "CHANGELOG-byte-grep discovery channel
  emits drop-with-reason span"
…onto runnable ground without depending on captureSwim() shim. The
span-emission piece is downstream (callsite needs to import this and
call emitContinuationDisabledSpan/equivalent with discovery.channel
attribute) but the discovery primitive is now testable in isolation.

Coverage (18 tests):
- subject-path: full subject, collision (multi-hit), empty subject
  (defense-in-depth: empty must NOT match-everything), whitespace,
  empty changelog, literal grep -F (no regex interpretation)
- PR-token extract: trailing #N, last-of-multiple, none, bare #N (must
  miss \u2014 only `(#N)` form), empty parens
- composite: load-bearing PR-channel hit on openclaw#70595, fallback to subject
  for harness-divergence instances, NO-hit for harness-divergence
  (memo says CHANGELOG silent for these)
- channel attribution: needle=PR-token vs needle=subject distinguishes
  channels for downstream span attrs

Boundary: pure string-in, structured-out. No git, no fs, no network.
Caller drives `git show <base>:CHANGELOG.md` and
`git log -1 --format='%s' <commit>`. Discipline matches the helper-
tier contract pattern from PR #406 (94fc8d1).

Result: 18/18 passing. swim-37 vitest project: 52 passed | 18 todo (4).
pnpm lint:core: 0/0.

Closes part of #324; wires §1 it.todo placeholder.

* feat(swim-37): #324 §1 cross-cut cherry-pick provenance parser

Adds the SECOND §1 discovery channel from the trap-class memo cross-cut
(L168 of studies/swim-37/traps/parallel-evolution-class.md).

Where CHANGELOG-byte-grep catches the rebase-source's WORK already in
base, cherry-pick-provenance catches the rebase-source's COMMIT already
in base — by parsing the `(cherry picked from commit <sha>)` footer
that git appends. The trap-class memo notes Instance 3 (`aa1908bf38`)
carries this footer pointing to `9dd097a7a5...`, which would have been
caught deterministically before conflict triage if the parser existed.

API:
- parseCherryPickProvenance(commitBody) — returns ALL footers in body
  order; multi-pick provenance accumulates when commits cross multiple
  branches; caller decides which is load-bearing
- lastCherryPickSourceSha(commitBody) — convenience for the common
  case (most recent pick in chain)

Boundary: pure parser. Caller drives `git show -s --format=%B` to
get body and `git merge-base --is-ancestor <src> <base>` to make the
ancestor decision.

Coverage (17 tests):
- byte-pinned aa1908b footer (40-hex)
- multi-footer in body order (cross-branch picks)
- trailing-period acceptance
- short-SHA acceptance (7+ hex, git's unique-floor)
- non-hex rejection, too-short rejection
- anchored-line discipline (sentence with phrase ≠ footer)
- leading-whitespace tolerance (indented footer blocks)
- empty body, no-footer body
- revert-footer NOT confused with pick-footer
- evidence preservation (line field for downstream span attrs)

Companion to chunk-1 of this PR (changelog-grep). Together they cover
the two positive-signal channels from the §1 memo. The conflict-content
classification rubric (third channel) is harder to cover as a pure
function — needs file-path + diff inspection — so it's left for a later
slice with the right shape.

Result: 17/17 passing this file. swim-37 vitest project: 69 passed |
18 todo (5 files). pnpm lint:core: 0/0.

Closes more of #324; second of two §1 discovery primitives.

* feat(swim-37): #324 §1 entry-point classifier (composes both discovery primitives)

Wires the §1 entry-point it.todo placeholder
  swim-runner.test.ts L80: "rebase bot classifies synthetic squash-rebased
  commit as DROP (not PICK)"
onto runnable ground by composing the two §1 discovery primitives
(changelog-grep + cherry-pick-provenance) into a single DROP/PICK/REVIEW
verdict with channel attribution and structured evidence for downstream
span emission.

API:
- classifyRebasePick({subject, commitBody, baseChangelog, isAncestorOf})
  returns {verdict: DROP|PICK|REVIEW, channel, evidence,
           needsConflictContentInspection?}

Channel ordering (memo-aligned, deterministic):
  1. cherry-pick-provenance (highest precision: exact source SHA,
     ancestor-of-base check unambiguous)
  2. changelog-grep:pr (high precision: PR tokens rarely collide)
  3. changelog-grep:subject (medium precision: substring, can collide)

Load-bearing invariant: when both positive-signal channels miss, the
verdict is REVIEW with needsConflictContentInspection=true \u2014 NOT a
silent default to PICK. The memo's third channel (conflict-content
classification rubric) needs file-path + diff inspection, not pure
string\u2192struct, so it's NOT covered here; classifier surfaces the gap
explicitly so the caller can hand off rather than misclassify the
test-harness divergence cases.

Evidence completeness: even when one channel is load-bearing for the
verdict, ALL channels' evidence is recorded in the result so the
downstream span emission can carry full audit data. Non-ancestor
cherry-pick footers go to a separate
slot for cross-branch provenance audit.

Boundary: pure function. Caller drives:
  - subject from `git log -1 --format='%s' <pick>`
  - body from `git show -s --format=%B <pick>`
  - base CHANGELOG from `git show <base>:CHANGELOG.md`
  - isAncestorOf wrapping `git merge-base --is-ancestor`

Coverage (10 tests):
- Instance 1 (openclaw#70595): DROP via changelog-grep:pr; no
  needsConflictContentInspection on DROP
- Instance 3 (cherry-pick): DROP via cherry-pick-provenance;
  precedence over CHANGELOG when both fire; non-ancestor footer
  doesn't trigger DROP via that channel
- Instance 2 (test-harness): REVIEW with
  needsConflictContentInspection=true; classifier never defaults to
  PICK on absence-of-signal
- subject-channel: DROP via changelog-grep:subject
- evidence completeness: both channels' evidence recorded when both
  fire; non-ancestor footers separated from load-bearing slot

Result: 10/10 passing this file. swim-37 vitest project: 79 passed |
18 todo (6 files). pnpm lint:core: 0/0.

Closes part of #324; third of three §1 todos wired to runnable ground.

---------

Co-authored-by: Ronan 🌊 <ronan@solidor.io>

Now that #408 (PR #408 merged at cb73bc8) landed
`rebase-classifier.ts` with the pure `classifyRebasePick` composer,
the integration-tier \u00a71 entry-point it.todo on swim-runner.test.ts
L80 is wireable without any new shim.

Synthetic commit pinned: subject mentions PR openclaw#70595, base CHANGELOG
contains the same PR token (squash-rebase shape from the memo's
7ee46a3 anchor instance). Discovery channel: changelog-grep:pr.

L81 (CHANGELOG-byte-grep emits drop-with-reason span) stays todo
\u2014 needs the production callsite that pairs the channel with
`emitContinuationDisabledSpan`, per \ud83c\udf0a's #408 note.

Test delta:
  Before: 20 passed | 16 todo (36)
  After: 21 passed | 15 todo (36)
  +1 live | -1 todo

…405)

* test(swim-37): wire captureSwim() for continue_work primitive

Replaces the placeholder `declare function captureSwim` in
`studies/swim-37/harness/swim-runner.test.ts` with a real implementation
in `studies/swim-37/harness/swim-runner.ts`. The wired path drives
`emitContinuationWorkSpan` against `createInMemorySpanRecorder()` and
returns the captured spans + the synthesized `chainId` (uuid v7 via
`generateChainId`).

Flips two prior `it.todo` markers to live tests:
  - emits continuation.work span with chain.id stamped (#366)
  - span carries chain.step.remaining attribute

Plus three new live tests:
  - sentinel: captureSwim() is wired for continue_work
  - refuses continue_delegate / heartbeat / lich primitives with a clear
    error so the spec's remaining `it.todo` markers stay honest about
    what is implemented
  - repeated calls don't leak capture state between invocations

STDOUT-only discipline preserved: no BasicTracerProvider, no OTLP
exporter, no @opentelemetry/sdk-trace-base machinery — capture flows
through `setContinuationTracer(recorder.tracer)` as documented in
README.md.

Other primitives (`continue_delegate`, `heartbeat`, lich-shape) remain
`it.todo` until the corresponding dispatch / heartbeat /
compaction-release seams have a comparable single-helper entry point we
can drive synthetically.

Test results:
  - swim-37 vitest project: 20 passed, 16 todo, 0 failed
  - in-memory-span-recorder.test.ts unchanged (12/12 still pass)
  - tsgo errors visible elsewhere in tree are pre-existing
    (src/plugin-sdk/provider-tools.ts), not introduced by this change

Refs #324 (swim-37 harness) — `it.todo` count drops from 17 \u2192 16,
first concrete primitive driven through the live tracer registry.

* docs(swim-37): continue_delegate wiring memo (memo-before-wire)

Pre-PR design memo for the next swim-37 primitive after #405
(continue_work). Decides shape before the wire so the PR lands clean.

Resolved by reading production code:
  Q1 — real setTimeout vs fake timers
       N/A for dispatch-accept (synchronous, pre-timer); banked for
       a separate continue_delegate_fire swim.

Cohort sign-off requested on:
  Q2 — recipient fan-out shape: N spans sharing chain.id (matches
       chunk-3 cohort design pin + #355 Stage-2 budget semantics).
  Q3 — delegate.mode x delegate.delivery: 8-cell it.each matrix
       + omission-contract row.

Banked for follow-up:
  Q4 — continue_delegate_fire as separate primitive, not sub-case;
       informs SwimPrimitive type-union shape.

Standard applied (per Cael 2026-04-27 #1498505918, figs
#1498505870): would skipping this memo cost a chunk's worth of
rework? Yes — three open Qs would have changed file shape.

Refs:
  - #405 (continue_work wired)
  - #324 (swim-37 harness)
  - docs/design/334-slice2-chunk5b-delegate-fire-memo.md (pattern)

* test(swim-37): wire captureSwim() for continue_delegate primitive

Implements the design from
`docs/design/swim-37-continue-delegate-wiring-memo.md` (commit
3bb086c762, same branch).

Wiring (swim-runner.ts):
- New continue_delegate case in the switch.
- Adds CaptureSwimOptions axes: recipients, delivery, delegateMode.
- Drives emitContinuationDelegateSpan once per recipient with shared
  chain.id (per chunk-3 cohort design pin: N recipients = N spans
  sharing chain.id, NOT one span with a recipients list).
- Validates recipients is a positive integer.

Tests (swim-runner.test.ts):
- 8-cell it.each matrix: delegate.mode (normal | silent | silent-wake
  | post-compaction) x delegate.delivery (immediate | timer).
- Omission contract: delegate.mode attribute absent when caller
  passes undefined.
- Fan-out: 3 recipients emit 3 spans with shared chain.id.
- Validation: recipients=0 and recipients=1.5 reject.
- Removes the prior 'refuses continue_delegate' test row (now wired).

Two it.todo markers preserved for downstream concerns:
- chain-budget decrement / ChainBudget.declineToCarry observable
  (needs ChainBudget integration, not in scope here).
- fan-out budget arithmetic per #355 Stage-2 (1 step not N) \u2014
  separate from span cardinality.

Test results delta:
  Before: 8 passed | 16 todo (24 tests)
  After: 19 passed | 15 todo (34 tests)
  +11 live | -1 todo

Refs:
- #324 swim-37 harness
- #405 captureSwim continue_work (parent of this branch)
- docs/design/swim-37-continue-delegate-wiring-memo.md

* test(swim-37): pin recipient.index gap on continue_delegate fan-out

Per Cael's caution on the wiring memo (Discord msg
1498507232185286849 sign-off): N spans sharing chain.id risks
collapsing into analytic mush at scale unless per-recipient
distinction is visible in attrs. Currently
emitContinuationDelegateSpan exposes no recipient.index axis \u2014
all N spans in a fan-out are byte-identical except spanId.

Adds:
- One live GAP-PIN test asserting the current (deficient) reality:
  fan-out spans toEqual on attributes. This will fail loudly when
  the production helper grows the axis, prompting a flip to
  not.toEqual.
- One it.todo marker for the upcoming production-helper change.

Production-helper change is out of scope for this PR; will file
follow-up issue. The gap-pin lives in the integration tier so it
shows up in any code-review touching delegate-dispatch span shape.

Refs:
- #324 swim-37 harness
- #405 captureSwim continue_work + continue_delegate (this PR)
- docs/design/swim-37-continue-delegate-wiring-memo.md (Q2)
- Cael Discord sign-off msg 1498507232185286849

* test(swim-37): wire \u00a71 entry-point todo onto classifyRebasePick

Now that #408 (PR #408 merged at cb73bc8) landed
`rebase-classifier.ts` with the pure `classifyRebasePick` composer,
the integration-tier \u00a71 entry-point it.todo on swim-runner.test.ts
L80 is wireable without any new shim.

Synthetic commit pinned: subject mentions PR openclaw#70595, base CHANGELOG
contains the same PR token (squash-rebase shape from the memo's
7ee46a3 anchor instance). Discovery channel: changelog-grep:pr.

L81 (CHANGELOG-byte-grep emits drop-with-reason span) stays todo
\u2014 needs the production callsite that pairs the channel with
`emitContinuationDisabledSpan`, per \ud83c\udf0a's #408 note.

Test delta:
  Before: 20 passed | 16 todo (36)
  After: 21 passed | 15 todo (36)
  +1 live | -1 todo

…el (#408)

* feat(swim-37): #324 §1 trap-class CHANGELOG-byte-grep discovery channel

Wires swim-37 trap-class §1 (parallel-evolution / cherry-false-negative)
discovery channel onto runnable ground. Pure function, byte-pinned to
studies/swim-37/traps/parallel-evolution-class.md §2 byte-walk against
v2026.4.24 CHANGELOG L164 (instance `7ee46a3ab9 fix: Add runner label
to /status (openclaw#70595)`).

API:
- grepChangelog(subject, content) — literal grep -F semantics, returns
  ALL hits (memo flags subject-line collisions as known false-positive
  mode; channel returns all so caller can decide)
- extractPrNumberToken(subject) — pulls trailing `(#N)` PR-token, last
  match wins (commits sometimes cite earlier PRs in body but subject's
  trailing PR is the landing PR)
- discoverChangelogHit(subject, content) — composite: PR-token first,
  fall through to full-subject; needle field tells caller which channel
  found the hit (or which was attempted last on miss)

Wires trap §1 it.todo:
  swim-runner.test.ts L81: "CHANGELOG-byte-grep discovery channel
  emits drop-with-reason span"
…onto runnable ground without depending on captureSwim() shim. The
span-emission piece is downstream (callsite needs to import this and
call emitContinuationDisabledSpan/equivalent with discovery.channel
attribute) but the discovery primitive is now testable in isolation.

Coverage (18 tests):
- subject-path: full subject, collision (multi-hit), empty subject
  (defense-in-depth: empty must NOT match-everything), whitespace,
  empty changelog, literal grep -F (no regex interpretation)
- PR-token extract: trailing #N, last-of-multiple, none, bare #N (must
  miss \u2014 only `(#N)` form), empty parens
- composite: load-bearing PR-channel hit on openclaw#70595, fallback to subject
  for harness-divergence instances, NO-hit for harness-divergence
  (memo says CHANGELOG silent for these)
- channel attribution: needle=PR-token vs needle=subject distinguishes
  channels for downstream span attrs

Boundary: pure string-in, structured-out. No git, no fs, no network.
Caller drives `git show <base>:CHANGELOG.md` and
`git log -1 --format='%s' <commit>`. Discipline matches the helper-
tier contract pattern from PR #406 (94fc8d1).

Result: 18/18 passing. swim-37 vitest project: 52 passed | 18 todo (4).
pnpm lint:core: 0/0.

Closes part of #324; wires §1 it.todo placeholder.

* feat(swim-37): #324 §1 cross-cut cherry-pick provenance parser

Adds the SECOND §1 discovery channel from the trap-class memo cross-cut
(L168 of studies/swim-37/traps/parallel-evolution-class.md).

Where CHANGELOG-byte-grep catches the rebase-source's WORK already in
base, cherry-pick-provenance catches the rebase-source's COMMIT already
in base — by parsing the `(cherry picked from commit <sha>)` footer
that git appends. The trap-class memo notes Instance 3 (`aa1908bf38`)
carries this footer pointing to `9dd097a7a5...`, which would have been
caught deterministically before conflict triage if the parser existed.

API:
- parseCherryPickProvenance(commitBody) — returns ALL footers in body
  order; multi-pick provenance accumulates when commits cross multiple
  branches; caller decides which is load-bearing
- lastCherryPickSourceSha(commitBody) — convenience for the common
  case (most recent pick in chain)

Boundary: pure parser. Caller drives `git show -s --format=%B` to
get body and `git merge-base --is-ancestor <src> <base>` to make the
ancestor decision.

Coverage (17 tests):
- byte-pinned aa1908b footer (40-hex)
- multi-footer in body order (cross-branch picks)
- trailing-period acceptance
- short-SHA acceptance (7+ hex, git's unique-floor)
- non-hex rejection, too-short rejection
- anchored-line discipline (sentence with phrase ≠ footer)
- leading-whitespace tolerance (indented footer blocks)
- empty body, no-footer body
- revert-footer NOT confused with pick-footer
- evidence preservation (line field for downstream span attrs)

Companion to chunk-1 of this PR (changelog-grep). Together they cover
the two positive-signal channels from the §1 memo. The conflict-content
classification rubric (third channel) is harder to cover as a pure
function — needs file-path + diff inspection — so it's left for a later
slice with the right shape.

Result: 17/17 passing this file. swim-37 vitest project: 69 passed |
18 todo (5 files). pnpm lint:core: 0/0.

Closes more of #324; second of two §1 discovery primitives.

* feat(swim-37): #324 §1 entry-point classifier (composes both discovery primitives)

Wires the §1 entry-point it.todo placeholder
  swim-runner.test.ts L80: "rebase bot classifies synthetic squash-rebased
  commit as DROP (not PICK)"
onto runnable ground by composing the two §1 discovery primitives
(changelog-grep + cherry-pick-provenance) into a single DROP/PICK/REVIEW
verdict with channel attribution and structured evidence for downstream
span emission.

API:
- classifyRebasePick({subject, commitBody, baseChangelog, isAncestorOf})
  returns {verdict: DROP|PICK|REVIEW, channel, evidence,
           needsConflictContentInspection?}

Channel ordering (memo-aligned, deterministic):
  1. cherry-pick-provenance (highest precision: exact source SHA,
     ancestor-of-base check unambiguous)
  2. changelog-grep:pr (high precision: PR tokens rarely collide)
  3. changelog-grep:subject (medium precision: substring, can collide)

Load-bearing invariant: when both positive-signal channels miss, the
verdict is REVIEW with needsConflictContentInspection=true \u2014 NOT a
silent default to PICK. The memo's third channel (conflict-content
classification rubric) needs file-path + diff inspection, not pure
string\u2192struct, so it's NOT covered here; classifier surfaces the gap
explicitly so the caller can hand off rather than misclassify the
test-harness divergence cases.

Evidence completeness: even when one channel is load-bearing for the
verdict, ALL channels' evidence is recorded in the result so the
downstream span emission can carry full audit data. Non-ancestor
cherry-pick footers go to a separate
slot for cross-branch provenance audit.

Boundary: pure function. Caller drives:
  - subject from `git log -1 --format='%s' <pick>`
  - body from `git show -s --format=%B <pick>`
  - base CHANGELOG from `git show <base>:CHANGELOG.md`
  - isAncestorOf wrapping `git merge-base --is-ancestor`

Coverage (10 tests):
- Instance 1 (openclaw#70595): DROP via changelog-grep:pr; no
  needsConflictContentInspection on DROP
- Instance 3 (cherry-pick): DROP via cherry-pick-provenance;
  precedence over CHANGELOG when both fire; non-ancestor footer
  doesn't trigger DROP via that channel
- Instance 2 (test-harness): REVIEW with
  needsConflictContentInspection=true; classifier never defaults to
  PICK on absence-of-signal
- subject-channel: DROP via changelog-grep:subject
- evidence completeness: both channels' evidence recorded when both
  fire; non-ancestor footers separated from load-bearing slot

Result: 10/10 passing this file. swim-37 vitest project: 79 passed |
18 todo (6 files). pnpm lint:core: 0/0.

Closes part of #324; third of three §1 todos wired to runnable ground.

---------

Co-authored-by: Ronan 🌊 <ronan@solidor.io>

…405)

* test(swim-37): wire captureSwim() for continue_work primitive

Replaces the placeholder `declare function captureSwim` in
`studies/swim-37/harness/swim-runner.test.ts` with a real implementation
in `studies/swim-37/harness/swim-runner.ts`. The wired path drives
`emitContinuationWorkSpan` against `createInMemorySpanRecorder()` and
returns the captured spans + the synthesized `chainId` (uuid v7 via
`generateChainId`).

Flips two prior `it.todo` markers to live tests:
  - emits continuation.work span with chain.id stamped (#366)
  - span carries chain.step.remaining attribute

Plus three new live tests:
  - sentinel: captureSwim() is wired for continue_work
  - refuses continue_delegate / heartbeat / lich primitives with a clear
    error so the spec's remaining `it.todo` markers stay honest about
    what is implemented
  - repeated calls don't leak capture state between invocations

STDOUT-only discipline preserved: no BasicTracerProvider, no OTLP
exporter, no @opentelemetry/sdk-trace-base machinery — capture flows
through `setContinuationTracer(recorder.tracer)` as documented in
README.md.

Other primitives (`continue_delegate`, `heartbeat`, lich-shape) remain
`it.todo` until the corresponding dispatch / heartbeat /
compaction-release seams have a comparable single-helper entry point we
can drive synthetically.

Test results:
  - swim-37 vitest project: 20 passed, 16 todo, 0 failed
  - in-memory-span-recorder.test.ts unchanged (12/12 still pass)
  - tsgo errors visible elsewhere in tree are pre-existing
    (src/plugin-sdk/provider-tools.ts), not introduced by this change

Refs #324 (swim-37 harness) — `it.todo` count drops from 17 \u2192 16,
first concrete primitive driven through the live tracer registry.

* docs(swim-37): continue_delegate wiring memo (memo-before-wire)

Pre-PR design memo for the next swim-37 primitive after #405
(continue_work). Decides shape before the wire so the PR lands clean.

Resolved by reading production code:
  Q1 — real setTimeout vs fake timers
       N/A for dispatch-accept (synchronous, pre-timer); banked for
       a separate continue_delegate_fire swim.

Cohort sign-off requested on:
  Q2 — recipient fan-out shape: N spans sharing chain.id (matches
       chunk-3 cohort design pin + #355 Stage-2 budget semantics).
  Q3 — delegate.mode x delegate.delivery: 8-cell it.each matrix
       + omission-contract row.

Banked for follow-up:
  Q4 — continue_delegate_fire as separate primitive, not sub-case;
       informs SwimPrimitive type-union shape.

Standard applied (per Cael 2026-04-27 #1498505918, figs
#1498505870): would skipping this memo cost a chunk's worth of
rework? Yes — three open Qs would have changed file shape.

Refs:
  - #405 (continue_work wired)
  - #324 (swim-37 harness)
  - docs/design/334-slice2-chunk5b-delegate-fire-memo.md (pattern)

* test(swim-37): wire captureSwim() for continue_delegate primitive

Implements the design from
`docs/design/swim-37-continue-delegate-wiring-memo.md` (commit
3bb086c762, same branch).

Wiring (swim-runner.ts):
- New continue_delegate case in the switch.
- Adds CaptureSwimOptions axes: recipients, delivery, delegateMode.
- Drives emitContinuationDelegateSpan once per recipient with shared
  chain.id (per chunk-3 cohort design pin: N recipients = N spans
  sharing chain.id, NOT one span with a recipients list).
- Validates recipients is a positive integer.

Tests (swim-runner.test.ts):
- 8-cell it.each matrix: delegate.mode (normal | silent | silent-wake
  | post-compaction) x delegate.delivery (immediate | timer).
- Omission contract: delegate.mode attribute absent when caller
  passes undefined.
- Fan-out: 3 recipients emit 3 spans with shared chain.id.
- Validation: recipients=0 and recipients=1.5 reject.
- Removes the prior 'refuses continue_delegate' test row (now wired).

Two it.todo markers preserved for downstream concerns:
- chain-budget decrement / ChainBudget.declineToCarry observable
  (needs ChainBudget integration, not in scope here).
- fan-out budget arithmetic per #355 Stage-2 (1 step not N) \u2014
  separate from span cardinality.

Test results delta:
  Before: 8 passed | 16 todo (24 tests)
  After: 19 passed | 15 todo (34 tests)
  +11 live | -1 todo

Refs:
- #324 swim-37 harness
- #405 captureSwim continue_work (parent of this branch)
- docs/design/swim-37-continue-delegate-wiring-memo.md

* test(swim-37): pin recipient.index gap on continue_delegate fan-out

Per Cael's caution on the wiring memo (Discord msg
1498507232185286849 sign-off): N spans sharing chain.id risks
collapsing into analytic mush at scale unless per-recipient
distinction is visible in attrs. Currently
emitContinuationDelegateSpan exposes no recipient.index axis \u2014
all N spans in a fan-out are byte-identical except spanId.

Adds:
- One live GAP-PIN test asserting the current (deficient) reality:
  fan-out spans toEqual on attributes. This will fail loudly when
  the production helper grows the axis, prompting a flip to
  not.toEqual.
- One it.todo marker for the upcoming production-helper change.

Production-helper change is out of scope for this PR; will file
follow-up issue. The gap-pin lives in the integration tier so it
shows up in any code-review touching delegate-dispatch span shape.

Refs:
- #324 swim-37 harness
- #405 captureSwim continue_work + continue_delegate (this PR)
- docs/design/swim-37-continue-delegate-wiring-memo.md (Q2)
- Cael Discord sign-off msg 1498507232185286849

* test(swim-37): wire \u00a71 entry-point todo onto classifyRebasePick

Now that #408 (PR #408 merged at cb73bc8) landed
`rebase-classifier.ts` with the pure `classifyRebasePick` composer,
the integration-tier \u00a71 entry-point it.todo on swim-runner.test.ts
L80 is wireable without any new shim.

Synthetic commit pinned: subject mentions PR openclaw#70595, base CHANGELOG
contains the same PR token (squash-rebase shape from the memo's
7ee46a3 anchor instance). Discovery channel: changelog-grep:pr.

L81 (CHANGELOG-byte-grep emits drop-with-reason span) stays todo
\u2014 needs the production callsite that pairs the channel with
`emitContinuationDisabledSpan`, per \ud83c\udf0a's #408 note.

Test delta:
  Before: 20 passed | 16 todo (36)
  After: 21 passed | 15 todo (36)
  +1 live | -1 todo

* Add runner label to status output

* Add changelog entry for status runner label

* Fix status runner detection and sanitization

* Add runner label to status output

* Add changelog entry for status runner label

* Fix status runner detection and sanitization