Skip to content

fix(acp): stop the shared effort overlay seeding new KAS sessions - #21376

Merged
chenmingwei23 merged 1 commit into
mainfrom
fix/effort-overlay-leak-21362
Oct 19, 2026
Merged

chenmingwei23 merged 1 commit into
mainfrom
fix/effort-overlay-leak-21362

Conversation

@iamwhatever

@iamwhatever iamwhatever commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: A new chat with no reasoning effort picked runs at the model's own default, not at a level some other chat in the same project picked.

A new KAS chat on claude-opus-5.5 with nothing pinned (no slot pick, crew pin, role or agent.reasoning_effort) runs at whatever level the last chat in the same work dir picked, often max, while the composer still shows "Default". The study in #21362 measured max against medium on one host: thinking step p50 6.2s → 9.1s and p99 34.6s → 73.0s, sidebar turn p50 42s → 100s, credits per turn p50 6.4 → 17.2.

Why it matters

KAS is the default backend, so most installs are hit. One max pick in a project silently makes every later new chat in that project slower and dearer, and the UI says "Default", so nothing tells the user why. It is the top driver of the "mate thinks too much" reports.

Not a goal

  • kiro-cli (acp_backend=kiro) work-dir sharing. kiro-cli reads cli.json itself, so that sharing is the engine's own design and stays as it is.
  • Cleaning effort entries older builds already left in users' cli.json files. KAS no longer reads them.
  • Changing the effort tier order or what the composer shows.

What changed (motivation → approach → change)

New KAS chats now start at the model's own default. The shared cli.json effort keys are read, written and cleared only for kiro-cli, the one engine that reads that file. A level picked live still reaches the session's restart and reset successors.

Why it happened: <work_dir>/.kiro/settings/cli.json is shared by every session in a project. Provider construction read its effort keys back for every member of ACP_BACKENDS_KIRO_SLASH_COMMANDS, which includes KAS. KAS never opens the file, so the read-back was its only effect, and it handed one chat's pick to the next one.

The fix: a new set, ACP_BACKENDS_EFFORT_OVERLAY (kiro-cli only), gates every effort read, write and clear of that file. It has the same shape as ACP_BACKENDS_TOOL_SEARCH_OVERLAY, which already keeps KAS out of the Tool Search keys of the same file.

The read-back was also how a level picked live reached a compaction restart or a reset successor: the dashboard effort route retired the session's recorded override and left the file to carry the pick. The route now records the picked level on the session instead (record_effort_override in session_parts/claims.py, bound on SessionManager in src/kiro_crew/session.py, replacing retire_effort_override; a level cleared to the default drops the record), so allocation_identity starts the successor on the level the session runs. On kiro-cli that is the level the file already holds. On claude and codex, which read no file, a successor used to start at the default and now keeps the pick too.

Site KAS before KAS after kiro-cli
provider init read-back seeded from file none same
_apply_effort_overlay wrote file no-op same
change_effort rollback cleared file none same
clear_effort, no default cleared file, reset reset only same
spawn receipt _effort_channel cli_json config_option same
successor after a live pick file read-back recorded pick recorded pick

A KAS slot keeps its own pick across a gateway restart through its slot metadata (metadata_codec._restore_model_fields, then reasoning_effort_override). A session the dashboard does not start (cron, Slack, webhook) resolves its level from config (crew pin, role, agent.reasoning_effort) on every cold start. It can still take a live pick from a dashboard slot surfaced from its channel; that pick now reaches its successors, but a cold start from the channel side reads no dashboard slot, for effort as for the model.

The set is registered where every set is: the disposition table in agent_sdk/backends.py, the acp/types.py and acp_backends.py re-exports, and agent_sdk/backend_cards.py (off the card, with its reason). The providers, session, harness-parity and harness-onboarding specs say the same, and the comment in dashboard/session_control_parts/model_switch.py now names kiro-cli.

Backwards compatibility

Compatible: nothing main accepts is refused, and on kiro-cli every read, write and clear of the file is unchanged. SessionManager.retire_effort_override is replaced by record_effort_override; its one caller, the dashboard effort route, changes with it.

KAS loses the file's read-back, write and clear. Every reader of a KAS session's effort, and of the removed method:

Reader: src/kiro_crew/providers/acp.py:AcpProvider.init -- chat -- a KAS slot's own pick arrives as effort_per_model, not from the file -- test_effort_overlay_readers.py::TestANewKasChatStartsAtTheModelDefault::test_a_level_the_slot_pinned_still_reaches_the_session
Reader: src/kiro_crew/dashboard/slot_persistence/metadata_codec.py:_restore_model_fields -- chat -- restores slot.reasoning_effort from slot metadata, never from cli.json -- test_slot_metadata_codec.py::test_every_field_a_purpose_reads_round_trips_and_every_other_is_left_alone
Reader: src/kiro_crew/dashboard/chat_api/slot_effort.py:api_chat_slot_reasoning_effort -- chat -- records the live pick on the session through record_effort_override -- test_chat_slot_reasoning_effort.py::TestChatSlotReasoningEffortLiveProvider::test_live_effort_capable_model_uses_change_effort_no_reset
Reader: src/kiro_crew/session_lifecycle_parts/parked_queue.py:allocation_identity -- chat -- a compaction or reset successor starts on the recorded live pick -- test_compaction_failure_pause.py::test_a_live_effort_pick_replaces_the_recorded_override
Reader: src/kiro_crew/config/loader.py:KiroCrewConfig.resolve_session_effort -- cron -- cron, Slack and webhook cold starts resolve from config, as before -- test_crew_reasoning_effort.py::TestFactoryAppliesTheCrewPin::test_crew_pin_reaches_the_provider_with_no_override_present
Reader: src/kiro_crew/providers/acp_model_effort.py:clear_effort -- chat -- still returns False so the handler resets, and the cold start reads no file -- test_effort_overlay_readers.py::TestKasLeavesTheSharedFileAlone::test_clearing_a_kas_level_asks_for_a_reset_and_leaves_the_file
Reader: src/kiro_crew/subagent_parts/model_choice.py:_effort_channel -- subagent -- a KAS spawn receipt names the effortLevel config option -- test_spawn_backend.py::TestEffortVerdict::test_the_applied_note_names_the_channel_the_backend_uses

Tests

  • test/test_effort_overlay_readers.py (new). A new KAS chat with another chat's max in the file resolves no level and pushes nothing. A slot's own pick still reaches KAS as effortLevel. KAS writes no file, and a refused live change or a clear leaves another harness's entry in place. kiro-cli still reads the file back and writes its own level.
  • test/test_compaction_failure_pause.py: a live pick replaces the recorded override, so the compaction successor starts on it; a live clear drops it.
  • test/test_chat_slot_reasoning_effort.py: the live route records the picked level ("" on a live clear), and records nothing on the deferred, reset and failed paths.
  • test/test_spawn_backend.py: a KAS spawn receipt names config option effortLevel.
  • test/test_harness_parity.py: each overlay writer names its own reader set; both sets leave KAS out and sit inside the slash dialect.
  • test/test_session_bindings.py: the member inventory names record_effort_override.

Red before: with the provider changes absent, 4 of the 7 reader tests fail; the headline is assert {'claude-opus-5.5': 'max'} == {}. With the session and route changes absent, the 4 new record checks fail.

Each guard put back to main's behaviour in a throwaway worktree, one at a time:

Mutation Caught by
init read-back gated on the slash dialect 2 tests, incl. the leak test
_apply_effort_overlay gated on the slash dialect 3 tests
change_effort rollback clear on the slash dialect the refused-change test
clear_effort clears for every kiro-family harness the clear test
_effort_channel reports cli_json for KAS the receipt test
the set admits KAS 5 tests, incl. the parity pin
the record retired instead of replaced the successor test
the route records no live pick 2 route tests
the route records the old slot level 2 route tests

Determinism, over the new and changed tests: 20 runs at -n0; 10 at -n 4 --dist loadgroup --max-worker-restart=2 beside test_acp_kas_effort.py and test_acp_tool_search_wire.py; one shuffled order, seed 11662. All green.

Local gates: black, isort, flake8, mypy (2777 files), comment-history, brand, harness-parity added-line gate, file-line cap, flake-pattern ratchet, repo semgrep rules, docs-lint. Related tests over 54 files: 4017 passed.

Manual verification

N/A — unit coverage sufficient: the leak lives in provider construction and the successor's level in the session record. The tests drive a real AcpProvider on a temporary work dir with a real cli.json, and a real SessionManager compaction restart. No live KAS session was started.

Screenshots / video

N/A — no frontend file changes.

Related Issues

Fixes #21362

Pattern harvest

Rule candidate: review-prompt
Pattern: per-session state persisted in a scope many sessions share (a work dir, a project file) and read back as one session's own value

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 216d0c07125e221071712119074c50e06019bf61 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Advisory — does not block merge.

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 16, 2026
@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 216d0c07125e221071712119074c50e06019bf61 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Before merging, check that a KAS slot's live effort pick survives a gateway restart through slot metadata alone, now that cli.json no longer carries it.

[DESIGN-REVIEWED] 216d0c0

Description read

[DESCRIPTION-READ] 0c8848fac120473b14a92362fe8e6a3b5aa528522be25b9603ed1b93cea857d2

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 216d0c07125e221071712119074c50e06019bf61 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

First-Principles-Verdict: PASS

Check that the metadata_codec._restore_model_fields → reasoning_effort_override path really replaces the deleted file read-back, which the old docstring credited with "a level set on a KAS session survives a gateway restart."

What this change ships

Inventory (7 items) — 7 justified

Intent: a new KAS chat with no effort picked runs at the model's own default, not at a level another chat in the same project left in the shared cli.json. FIX.

  1. A new KAS chat is no longer started from the work dir's shared cli.json effort entry — justified
  2. A KAS session no longer writes its effort into the shared cli.json, so a kiro-cli session in the same project no longer picks it up — justified
  3. Rolling back or clearing a KAS effort level no longer touches that file — justified
  4. A level picked live now carries over to the compaction or reset successor (record_effort_override replaces retire_effort_override) — justified
  5. On claude/codex, a reset successor now keeps the live pick instead of going back to the default — justified
  6. A KAS spawn receipt now names the effortLevel config option instead of cli.json — justified
  7. New capability set ACP_BACKENDS_EFFORT_OVERLAY = {kiro}, read at 5 sites (provider init, _apply_effort_overlay, change_effort rollback, clear_effort, _effort_channel), plus a backend-card off-card entry — justified

[FIRST-PRINCIPLES-REVIEWED] 216d0c0

Description read

[DESCRIPTION-READ] 6fbc2ddeff1f3bb771bb6988f074a9aa56950a771aa9c22d3c2fbd487dddd286

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 216d0c07125e221071712119074c50e06019bf61 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 216d0c0

Verdict parsed from the review's SHA-scoped output markers for commit 216d0c07125e221071712119074c50e06019bf61.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 216d0c07125e221071712119074c50e06019bf61: <one-sentence reason>

@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 216d0c07125e221071712119074c50e06019bf61 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 216d0c0

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 216d0c07125e221071712119074c50e06019bf61: <one-sentence reason>

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • span=2c1ef3ea2972 _effort_channel(None) names cli_json on a default-KAS install: rebutted (disproportional, unchanged by this diff)

The None arm answers the same as on main: there, a spawn with no backend already got a cli_json receipt while KAS took its level as effortLevel. This PR neither adds nor widens that gap.
Fixing it means resolving the gateway's default backend inside a reporting helper (resolve_selected_backend with its KAS-availability floor, member-DM routing, a shared subagent's parent runtime) and rewriting every backend=None receipt the spawn tests pin. That is new scope beyond chat effort.
The receipt is reporting-only: no level is applied differently because of it.
What this PR keeps true is _effort_channel's own contract for an explicit backend="kas", now config_option, pinned by test_spawn_backend.py::TestEffortVerdict::test_the_applied_note_names_the_channel_the_backend_uses.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • A KAS level set live comes back at that level after a gateway restart: checked, it holds

Pick: slot_effort.py commits slot.reasoning_effort = effort on the live, deferred and reset paths (test_chat_slot_reasoning_effort.py::TestChatSlotReasoningEffortLiveProvider::test_live_effort_capable_model_uses_change_effort_no_reset and ::test_active_turn_defers_live_push).
Persist and restore: the slot metadata codec writes reasoning_effort, and _restore_model_fields (metadata_codec.py:410) reads it back (test_slot_metadata_codec.py::test_every_field_a_purpose_reads_round_trips_and_every_other_is_left_alone).
Next turn: chat_runner hands it to get_or_create as reasoning_effort_override=slot.reasoning_effort or None (both _eager_spawn and the turn's _allocation_kwargs), and the factory keys it into effort_per_model (test_effort.py::TestFactoryEffortThreading).
KAS start: _apply_initial_effort pushes it as effortLevel even with another chat's max in cli.json (test_effort_overlay_readers.py::TestANewKasChatStartsAtTheModelDefault::test_a_level_the_slot_pinned_still_reaches_the_session). A warm-pool claim applies it post-claim through change_effort, which on KAS is a live push with no file write.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Design punchline (head db1e8cbe69): a KAS effort change from the dropdown writes slot.reasoning_effort. Checked, it does.

  • The dropdown route (dashboard/chat_api/slot_effort.py) sets slot.reasoning_effort = effort on all three success paths: after a live change_effort, on the deferred path while a turn runs, and before the reset fallback. test_chat_slot_reasoning_effort.py::TestChatSlotReasoningEffortLiveProvider::test_live_effort_capable_model_uses_change_effort_no_reset and ::test_active_turn_defers_live_push assert the slot value.
  • From there the level survives a restart through slot metadata alone. The chain, link by link with its tests, is in the First Principles disposition on this head.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Oct 16, 2026
chenmingwei23
chenmingwei23 previously approved these changes Oct 17, 2026

@chenmingwei23 chenmingwei23 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Approved via bulk merge run: no failing checks on the current revision.

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 17, 2026
@iamwhatever
iamwhatever force-pushed the fix/effort-overlay-leak-21362 branch from db1e8cb to 95e802b Compare October 17, 2026 04:47
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 17, 2026
@iamwhatever
iamwhatever force-pushed the fix/effort-overlay-leak-21362 branch from 95e802b to eaa62f6 Compare October 17, 2026 06:13
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: action required A blocking check or review needs attention labels Oct 17, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: no
mechanism: the dashboard effort route records a live pick on the session (record_effort_override) instead of retiring the record

  • "Channel sessions can never change effort live": the claim was wrong; fixed in eaa62f6

A dashboard slot surfaced from a channel can change that session's effort live (slot_effort.py pushes to effective_session_key(slot)).
Checking it exposed a wider gap: after any live pick, a compaction restart or reset successor started from allocation_identity, and the route had retired the recorded override so the cli.json read-back would carry the pick. With KAS out of the read-back that left the successor at the default.
The route now records the picked level (record_effort_override; "" drops it), so the successor starts on the level the session runs, on every harness. Pinned by test_compaction_failure_pause.py::test_a_live_effort_pick_replaces_the_recorded_override and ::test_a_live_clear_drops_the_recorded_override, mutation-proven.
Left as is: a channel-side cold start (gateway restart, idle reap) reads no dashboard slot, for effort as for the model (slack/handler.py, messaging/dispatch.py pass neither). The body now says so.

@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • A KAS live pick is kept on every route without cli.json: checked, one gap found and fixed in eaa62f6

Dropdown: slot_effort.py commits slot.reasoning_effort on the live, deferred and reset paths, and slot metadata restores it after a gateway restart.
/effort typed in a KAS chat: the engine refuses it (-32603, agent_sdk/backends.py), so no level changes and there is nothing to save.
Crew switch: the slot keeps its level, and the next cold start hands it to the factory as reasoning_effort_override, above the crew pin.
The gap: a compaction restart or reset successor took no slot level and leaned on the file. The route now records the live pick on the session (record_effort_override), so the successor starts on it; mutation-proven in test_compaction_failure_pause.py and test_chat_slot_reasoning_effort.py.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: checking Automated validation is still running labels Oct 17, 2026
@iamwhatever
iamwhatever force-pushed the fix/effort-overlay-leak-21362 branch from a8ce2f9 to da41dde Compare October 18, 2026 07:11
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Rebased onto main 9d088c8 (new head da41dde, still one commit).

Conflicted file: src/kiro_crew/agent_sdk/backends.py (2 hunks). Main split this module into backends_* submodules and turned backends.py into a re-export hub. Resolution keeps main's split as-is and moves this PR's two edits to where the code now lives:

  • ACP_BACKENDS_EFFORT_OVERLAY = frozenset({ACP_BACKEND_KIRO}) and its comment: added in backends_mcp.py next to ACP_BACKENDS_TOOL_SEARCH_OVERLAY, and re-exported from backends.py (the doc-table row in backends.py auto-merged).
  • The ACP_BACKENDS_KIRO_SLASH_COMMANDS comment (slash dialect no longer decides who reads the cli.json effort keys): applied in backends_turns.py.

Nothing from main was dropped: a def symbol diff of every touched .py vs origin/main shows only retire_effort_override and test_a_live_effort_pick_retires_the_recorded_override gone, both removed by this PR on purpose (replaced by record_effort_override). Auto-merged without conflict: session.py, subagent_parts/model_choice.py, backend_cards.py, session.md, test_harness_parity.py, test_session_bindings.py.

Real Adapter Contract: the red on a8ce2f9 was test_the_installed_adapter_still_builds_the_frames_the_refusal_reads (codex-acp 2.2.2 no longer has createMcpRawInput). That is main's, fixed by #21581; not touched by this PR.

Local: 2,462 tests passed across the touched and adjacent files (effort, overlay, harness parity, capability sets, session bindings, spawn backend, backends vocabulary).

@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Oct 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/effort-overlay-leak-21362 branch from da41dde to aa3408e Compare October 18, 2026 07:38
@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: checking Automated validation is still running labels Oct 18, 2026
KAS never opens <work_dir>/.kiro/settings/cli.json, yet provider
construction read its effort keys back for every kiro-family session.
The file is keyed by work dir, so the level one chat picked became the
starting level of every later chat on that model in the project, while
the composer still showed Default.

Every effort read, write and clear of that file is now scoped to a new
ACP_BACKENDS_EFFORT_OVERLAY set (kiro-cli only), the same shape as
ACP_BACKENDS_TOOL_SEARCH_OVERLAY. A KAS slot keeps its own pick across
a gateway restart through its slot metadata.

The read-back was also how a level picked live reached a compaction
restart or reset successor: the dashboard effort route retired the
session's recorded override and left the file to carry the pick. It
now records the picked level instead (record_effort_override), so the
successor starts on the level the session runs on every harness.

Fixes #21362
@iamwhatever
iamwhatever force-pushed the fix/effort-overlay-leak-21362 branch from aa3408e to 216d0c0 Compare October 19, 2026 02:57
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: passed Eligible automated validation passed for the current revision labels Oct 19, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Rebased onto main 574cb18849 (from 943a4e5c0e). Still one commit: 216d0c0712. Size unchanged: 24 files, +284/-86. Every src/ and test/ hunk is the same as on aa3408e663; git range-diff shows changes only in the two docs below.

Conflicts. Both were one long table row where main and this PR edited different words. Each merged row keeps every edit from both sides.

Auto-merged, checked by hand: dashboard/chat_api/slot_effort.py (the #21799 switch notice sits in the reset branch; the record_effort_override call from this PR sits in the live-push branch), agent_sdk/backends_mcp.py, agent_sdk/backends_turns.py, test/test_harness_parity.py, harness-onboarding.md. The main edits there are path renames from the owner splits. Main has no new caller of retire_effort_override.

Local checks on the rebased commit: 20 related test files, 1325 passed / 5 skipped. On the final base, the touched tests plus parity, docs and facade tests: 910 passed. isort, flake8, mypy (changed src), scoped black, harness parity, comment history, line cap and docs-lint pass.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 19, 2026
@chenmingwei23
chenmingwei23 merged commit f161e27 into main Oct 19, 2026
103 of 104 checks passed
@chenmingwei23
chenmingwei23 deleted the fix/effort-overlay-leak-21362 branch October 19, 2026 14:11
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 19, 2026
rnoack1 pushed a commit to rnoack1/KiroCrew that referenced this pull request Oct 19, 2026
kirodotdev#22056's ratchet recorded sites kirodotdev#21376 removed 75s later. Three claude
captures from kirodotdev#21753 omit the agent_name kirodotdev#22064 then began requiring.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

reasoning effort leaks across chats: new session inherits work-dir overlay instead of default

2 participants