v0.125.1

Plugin manifests now list Bilko Bibitkov as owner and author, and the dev plugin moves to 0.2.2 so installed copies refresh. A new guard checks commits, pushes and CI so only the persona appears in the repo.

Also carries the web remote connection fixes already live on bilko.run.

Linux installers are not in this release, because the build runner changed and the job failed. They return with the next release. On Linux, use npx claude-code-session-manager@latest.

Earlier npm versions and GitHub releases are unchanged and are not being unpublished.

v0.125.0

Release v0.125.0

v0.124.1

Release v0.124.1

v0.123.1

Release v0.123.1

v0.123.0

Release v0.123.0

v0.122.1

Release v0.122.1

0.122.0

chore(release): bump to v0.122.0

- test(chat): drop dead SM_CHAT_CONCURRENCY handling from chat-queue spec
- refactor(scheduler): drop dead scheduledFor from queueStore
- test(chat): rewrite chat-queue spec against sessionSlots gating
- fix(scheduler): slots stepper max falls back to 10, not current cap
- test(scheduler): derive effectiveConcurrency clamp ceiling from MAX_SLOTS
- test(atomic-write): assert invariant, not wrapper spelling
- perf(classify): spawn classifyPromptTicket with --strict-mcp-config
- docs(slots): update SM_SESSION_SLOTS tip and chatRunner code-map line
- fix(history): loop-efficiency weekly usage reads ok-stale cache
- fix(scheduler): slot editors read ceiling from snapshot max
- refactor(scheduler): delete legacy global queue migration
- refactor(scheduler): delete on-reset fire timer, normalize persisted on-reset to when-available
- refactor(scheduler): drop on-reset firePolicy and scheduledFor from renderer/preload types
- feat(scheduler): raise session slots ceiling to 32
- perf(audit-log): sync rename + async streamed gzip roll, 32 MB size trigger
- feat(scheduler): stamp memory fields on every heartbeat line
- feat(chat): evict Epic chats on delete + LRU beyond 8, cap events at 500
- refactor(scheduler): replace SM_REVERIFY_PERIODIC_DISABLE env with test seam
- feat(history): loop-efficiency panel on History dashboard
- perf(model-info): tail-read history.jsonl instead of full parse
- perf(voice): lazy-load speech model on first record, release after 30m idle
- test(ipc): align firePolicy spec with on-reset removal
- fix(boot): run first-boot seeders serially
- refactor(scheduler): delete dead env kill switches and trivial wrappers
- feat(history): loop-efficiency IPC feeding computeLoopEfficiency
- fix(seed-dev-plugin): repair stale or missing marketplace path despite done marker
- docs(develop): make estimateMinutes optional in planner docs; PRD_AUTHORING v9
- refactor(scheduler): drop on-reset from firePolicy IPC schema
- refactor(scheduler): delete dead notifyNeedsReview and SM_REVIEW_NOTICE_IMMEDIATE
- feat(scheduler): add loopEfficiency calculator
- feat(mcp): make estimateMinutes optional in scheduler_create_prd schema
- fix(readiness): verify dev plugin install on disk (SKILL.md + marketplace dir)
- fix(seed): report failed after first failed attempt and show Plugins banner
- feat(launch): include persona body in every launch path's AIM prompt
- fix(chat): drop chatRunner private concurrency cap; gate only on sessionSlots
- perf(claude): one-shot claude -p helpers skip MCP servers via --strict-mcp-config
- perf(scheduler): emit --strict-mcp-config for persona scopes with no mcp__ tool
- docs(reviews): scheduler.cjs deletion audit at c429638
- docs(scheduler): reliability target and feature freeze
- feat(prd): make estimateMinutes optional on PRD create path
- docs(prd-authoring): drop transcript_errors workarounds, bump to v8
- refactor(dev-lead): drop verifier-trap rules obsoleted by outcome-only verify
- fix(scheduler): orphan success requires landed commit evidence
- feat(queue): make estimateMinutes optional in PRD lint
- feat(scheduler): cap post-run downgrade re-fires at 2, park as refire_capped
- refactor(runVerify): retire transcript pattern scan; verdict depends on outcomes only
- docs(manual): documents app v0.121.0 (no content change)
- feat(scheduler): refireCap helper for verify-downgrade re-fire cap
- fix(verify): checkDeps uses queue dependsOn only, drop prose/soak deps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

0.122.0

chore(release): bump to v0.122.0

- test(chat): drop dead SM_CHAT_CONCURRENCY handling from chat-queue spec
- refactor(scheduler): drop dead scheduledFor from queueStore
- test(chat): rewrite chat-queue spec against sessionSlots gating
- fix(scheduler): slots stepper max falls back to 10, not current cap
- test(scheduler): derive effectiveConcurrency clamp ceiling from MAX_SLOTS
- test(atomic-write): assert invariant, not wrapper spelling
- perf(classify): spawn classifyPromptTicket with --strict-mcp-config
- docs(slots): update SM_SESSION_SLOTS tip and chatRunner code-map line
- fix(history): loop-efficiency weekly usage reads ok-stale cache
- fix(scheduler): slot editors read ceiling from snapshot max
- refactor(scheduler): delete legacy global queue migration
- refactor(scheduler): delete on-reset fire timer, normalize persisted on-reset to when-available
- refactor(scheduler): drop on-reset firePolicy and scheduledFor from renderer/preload types
- feat(scheduler): raise session slots ceiling to 32
- perf(audit-log): sync rename + async streamed gzip roll, 32 MB size trigger
- feat(scheduler): stamp memory fields on every heartbeat line
- feat(chat): evict Epic chats on delete + LRU beyond 8, cap events at 500
- refactor(scheduler): replace SM_REVERIFY_PERIODIC_DISABLE env with test seam
- feat(history): loop-efficiency panel on History dashboard
- perf(model-info): tail-read history.jsonl instead of full parse
- perf(voice): lazy-load speech model on first record, release after 30m idle
- test(ipc): align firePolicy spec with on-reset removal
- fix(boot): run first-boot seeders serially
- refactor(scheduler): delete dead env kill switches and trivial wrappers
- feat(history): loop-efficiency IPC feeding computeLoopEfficiency
- fix(seed-dev-plugin): repair stale or missing marketplace path despite done marker
- docs(develop): make estimateMinutes optional in planner docs; PRD_AUTHORING v9
- refactor(scheduler): drop on-reset from firePolicy IPC schema
- refactor(scheduler): delete dead notifyNeedsReview and SM_REVIEW_NOTICE_IMMEDIATE
- feat(scheduler): add loopEfficiency calculator
- feat(mcp): make estimateMinutes optional in scheduler_create_prd schema
- fix(readiness): verify dev plugin install on disk (SKILL.md + marketplace dir)
- fix(seed): report failed after first failed attempt and show Plugins banner
- feat(launch): include persona body in every launch path's AIM prompt
- fix(chat): drop chatRunner private concurrency cap; gate only on sessionSlots
- perf(claude): one-shot claude -p helpers skip MCP servers via --strict-mcp-config
- perf(scheduler): emit --strict-mcp-config for persona scopes with no mcp__ tool
- docs(reviews): scheduler.cjs deletion audit at c429638
- docs(scheduler): reliability target and feature freeze
- feat(prd): make estimateMinutes optional on PRD create path
- docs(prd-authoring): drop transcript_errors workarounds, bump to v8
- refactor(dev-lead): drop verifier-trap rules obsoleted by outcome-only verify
- fix(scheduler): orphan success requires landed commit evidence
- feat(queue): make estimateMinutes optional in PRD lint
- feat(scheduler): cap post-run downgrade re-fires at 2, park as refire_capped
- refactor(runVerify): retire transcript pattern scan; verdict depends on outcomes only
- docs(manual): documents app v0.121.0 (no content change)
- feat(scheduler): refireCap helper for verify-downgrade re-fire cap
- fix(verify): checkDeps uses queue dependsOn only, drop prose/soak deps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

0.121.0

chore(release): bump to v0.121.0

- test(remote): align chatRunner and webRemoteDrilldown specs with shipped send/transcript contracts
- docs(validation): release-gate-green record — 2119 refuted, nothing landed
- docs(validation): lint-green-at-head review record
- test(web-remote): behavioural no-summary-push test replaces source-text smell
- chore(repo): untrack stray <path> sqlite MCP placeholder and ignore it
- docs(validation): sweep record for fix-phone-drop-summary-state
- docs(validation): phone read-and-reply plan record
- fix(prd): refuse duplicate bare slug within an Epic before NN allocation
- refactor(remote-app): drop per-tab summary state, listener and type
- feat(web-remote): cmd:chat:send returns send disposition
- chore(phone): delete dead SummaryPane component and its telemetry test
- feat(chat): enqueueExternalPrompt returns renderer disposition via ack registry
- docs(validation): macro inputs every press plan record
- feat(remote-app): macro input fields in NewSessionSheet, send macroId + inputs
- feat(web-remote): render HOT KEY macro inputs in cmd:session:create
- fix(macro): renderer renderMacroPrompt refuses undeclared placeholders, own-property reads, parity test
- fix(macro): saveMacro rejects prompt placeholders missing from merged inputs
- feat(macro): strict main-process macro prompt renderer and macroVisibleIn
- fix(remote): watch Main by tab sessionId, re-resolve late transcripts, drop per-turn summary push
- feat(remote-app): SessionScreen as readable chat with paging, pending bubble and polling
- feat(renderer): send returns ChatSendResult and onExternalSend reports it to main
- feat(remote): cmd:session:transcript serves paged conversation and resolves Main by tab sessionId
- feat(remote-app): paged transcript and typed send result in remoteApi
- feat(preload): add chat.reportExternalSendResult and ChatExternalSendResult contract
- feat(remote): add pure buildRemoteConversation for phone chat pages
- docs(manual): documents app v0.120.0 (no content change)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

0.120.0

chore(release): bump to v0.120.0

Covers 100 commits since v0.119.1: feat(chatRunner) ops error codes, feat(telemetry) unclean-shutdown classification, feat(tags) art + portfolio WorkTypes, feat(scheduler) persona tool scope, feat(retention) ops-retention settings, feat(health) dep-link checks, feat(remote-app) telemetry, plus fixes.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>