v0.3.0

Image

• image: ghcr.io/c360studio/semteams:0.3.0
• digest: sha256:7894207684e3d302a98db2d66b2fefc35b73796d5e491f4f606c2c1b5f4ac355

docker pull ghcr.io/c360studio/semteams:0.3.0

What's Changed

• ci(validate): add read-only local task validate mirroring Repository CI (#310) by @cglusky in #338
• fix(shell): replace remaining SemStreams identity strings in cmd/semteams (#308) by @cglusky in #339
• chore(runtime): resolve #283 foundation review follow-up nits (#303) by @cglusky in #342
• chore(runtime): resolve reviewer nits on #289, #293, and #295 (#302) by @cglusky in #343
• harness(agents): upgrade agent contracts, rules-index, workflow skills, Gemini routing, and parity suite (#340) by @cglusky in #341
• fix(semsource): implement message.Payload on StatusPayload (#285) by @cglusky in #346
• feat(configs): declare model endpoint pricing so agent.loop.cost-usd is stamped (#248) by @cglusky in #347
• feat(scripts): bring Taskfile and harness scripts to SemEngine parity (#344) by @cglusky in #345
• feat(pulse): make Program Pulse scheduled, incremental, and replay-safe (#268) by @cglusky in #348

Full Changelog: v0.2.0...v0.3.0

v0.2.0

Image

• image: ghcr.io/c360studio/semteams:0.2.0
• digest: sha256:30221cc26533d717774be6672feb595c6af3dd5c700e9aab4624a1be2af97242

docker pull ghcr.io/c360studio/semteams:0.2.0

What's Changed

• research: bound live pack loops with frozen spawn budgets by @cglusky in #311
• docs(agents): port .agents/ protocol, Gemini/Antigravity routing, and <24 KB AGENTS.md cap (#313) by @cglusky in #316
• work: structured baton, review, and decision overlays across portfolio repositories (#314) by @cglusky in #317
• ui(work): unify Project Work Board with inline Enjoy-style agentic drawer and scoped chat (#315) by @cglusky in #318
• ui(work): implement live GitHub WorkSource with baton, review, and decision gate derivation (#319) by @cglusky in #321
• ui(work): link live agentic runs to work items via [Work item: ...] prompt scope and run.issue.ref (#320) by @cglusky in #322
• test(work): cover inline run workbench, ChatBar context, and work_item_ref in work-board.spec.ts (#323) by @cglusky in #325
• feat(work): hydrate GitHub Projects v2 status options, pm_status, and priority (#324) by @cglusky in #326
• feat(tools): add portfolio-scoped github_read evidence adapter (#273) by @cglusky in #327
• chore(demo): retain autoresearch as opt-in framework demo outside MVP gate by @cglusky in #328
• fix(e2e): bound BuildKit cache and replace --no-cache with SOURCE_HASH invalidation by @cglusky in #329
• fix(e2e): re-point approval-pause and approval-resume onto live research pack by @cglusky in #330
• feat(research): evidence contract and mechanical derivation check (#271) by @cglusky in #334
• feat(research): complete containment budget — planner fan-out ceiling, cron cooldown/inflight, and kill switch (#272) by @cglusky in #336
• feat(pulse): fold Program Pulse (#267) into research pack with github_read and /pulse by @cglusky in #337

Full Changelog: v0.1.0...v0.2.0

v0.1.0

Image

• image: ghcr.io/c360studio/semteams:0.1.0
• digest: sha256:741499c713767000b73a3ac5d0d8003c635c9eb10f4890cf97c7c46dd45719f5

docker pull ghcr.io/c360studio/semteams:0.1.0

What's Changed

• fix(schemas): regenerate agentic-dispatch schema to match Go config by @cglusky in #4
• chore(ui): import semstreams-ui as ui/ subtree + wire into Taskfile, CI, lint by @cglusky in #1
• chore(ui): rename to @c360/semteams-ui and de-fiction the tree by @cglusky in #2
• chore(ui): remove deprecated "DO NOT EDIT" test file headers by @cglusky in #5
• docs(journeys): Playwright-first journey index (+ Phase C.1 pivot) by @cglusky in #3
• docs(ui): de-fiction residual multi-backend framing in ui/docs by @cglusky in #6
• feat(mock-llm): add YAML fixture loader for deterministic journey responses by @cglusky in #7
• feat(journeys): tool-approval-gate Playwright spec + agentic E2E stack by @cglusky in #8
• feat(journeys): real-time-activity-stream Playwright spec by @cglusky in #9
• feat(journeys): e2e:superpowers meta-runner + deliberate-break verification by @cglusky in #10
• fix(agentic-loop): keep test context alive across startLoopComponent helper by @cglusky in #11
• fix(tests): harden integration tests for macOS Docker Desktop by @cglusky in #12
• feat(ui): kanban board redesign — Phase 1 MVP by @cglusky in #13
• refactor: convert semteams from fork to import of semstreams by @cglusky in #14
• fix(tests): relax timing assertion in command registration attack test by @cglusky in #16
• feat(config): add .env.example + multi-provider agentic config by @cglusky in #15
• fix(ci): update container image name from semstreams to semteams by @cglusky in #17
• feat(workflow): add Deep Research as first real agentic workflow by @cglusky in #18
• fix(core): selective component registration + dispatch permissions by @cglusky in #19
• chore(ui): nuke legacy flow-editor E2E, promote journey config by @cglusky in #28
• test(ui): five new journey specs for the redesigned UI surfaces by @cglusky in #29
• fix(configs): rewrite e2e-agentic.json; skip tool-approval-gate (upstream gap) by @cglusky in #30
• chore(deps): bump semstreams v1.0.0-beta.15 → beta.24 by @cglusky in #31
• feat(approval): wire approval-flow UI end-to-end with X-User-Id seam by @cglusky in #32
• chore(deps): bump semstreams v1.0.0-beta.24 → beta.25 by @cglusky in #33
• docs(research): ADR-031 + open-questions memo for SemTeams research flow by @cglusky in #34
• feat(research): R1 — iterative researcher with reviewer-as-enumerator gate by @cglusky in #35
• feat(r2): add_source_repo tool — slice 1 of R2 source acquisition by @cglusky in #36
• feat(r2): source-acquisition e2e journey — slice 2 of R2 by @cglusky in #37
• feat(r2.5): chained research + source-acquisition journey by @cglusky in #38
• feat(r3.1): research.Artifact payload + ADR-031 stabilisation addendum by @cglusky in #39
• feat(r3.2.1): emit_research_artifact tool by @cglusky in #40
• feat(r3.2.2): mode-transition machinery — stabilisation + persona-swap by @cglusky in #41
• feat(r3.3): dev-via-spec port — planner / reviewer / challenger / architect-light chain by @cglusky in #42
• feat(r3.4): OSH dev-via-spec arc — chain converged on real LLM, spec emitted by @cglusky in #43
• docs(adr-032): R3.6 builder sandbox design — proposed (planning only, no code) by @cglusky in #44
• feat(r3.6.1): builder sandbox primitive (ADR-032) by @cglusky in #45
• feat(r3.6.1.1): conform sandbox to upstream wire format + read_only_paths by @cglusky in #46
• feat(r3.6.2.a): wire BashExecutor to sandbox via SANDBOX_URL by @cglusky in #47
• feat(r3.6.2.b): builder_decide terminal validator per ADR-032 §15 by @cglusky in #48
• feat(r3.6.2.c): dev-via-spec-builder persona fragments per ADR-032 §R3.6.2 by @cglusky in #49
• feat(r3.6.2.d): builder spawn rule + bootstrap_workspace tool by @cglusky in #50
• feat(r3.6.2.e): mock-LLM fixture extends dev-via-spec chain to builder by @cglusky in #51
• docs(r3.6.2.f): smoke #6 outcome + calibration findings by @cglusky in #52
• feat(r3.6.2.g): apply calibration + smoke #6 confirms tests_passing by @cglusky in #53
• docs(adr-033): harness-anchored verification + coordinator-as-el-jefe (Proposed) by @cglusky in #54
• feat(r3.7.1): harness catalog primitive (ADR-033) by @cglusky in #55
• feat(r3.7.2.a): verification.Commitment foundational primitive (ADR-033) by @cglusky in #56
• feat(r3.7.2.b): dev_via_spec.artifact carries verification_commitments[] by @cglusky in #57
• feat(r3.7.2.c): upstream-role verifiability discipline (persona-only) by @cglusky in #58
• docs(adr-034): qa-runner pattern adoption (proposed) by @cglusky in #60
• R3.7.2.e′: meshtasticd-3.x catalog entry + compose_profile relaxation by @cglusky in #62
• R3.7.2.f′: architect commitment contract by @cglusky in #63
• R3.7.2.g′: brownfield discovery walk (architect persona) by @cglusky in #64
• R3.7.2.h′: builder commitment-driven authoring + harness projection by @cglusky in #65
• R3.7.2.i′: evidence-rule registry + three concrete checker kinds by @cglusky in #66
• R3.7.2.j′: qa-reviewer persona + evidence.Summarize helper by @cglusky in #67
• R3.7.2.k′: mock-LLM e2e for builder → qa-reviewer chain shape by @cglusky in #68
• R3.7.2.l′: real-LLM smoke #7 prep — osh-demo wiring + playbook by @cglusky in #69
• R3.7.2.l′-runner: one-task smoke7 runner with capture + spend cap by @cglusky in #70
• fix(r3.7.2.l′-runner): single-quote smoke7:run echo to escape YAML # comment by @cglusky in #71
• feat(smoke7-runner): F3 chain-wedge report + tool-issues capture by @cglusky in #72
• chore: bump semstreams v1.0.0-beta.39 → v1.0.0-beta.43 by @cglusky in #73
• test(contract): pin dispatch default_tools wire shape (F4b regression guard) by @cglusky in #75
• chore: bump semstreams v1.0.0-beta.43 → v1.0.0-beta.45 (supersedes #74) by @cglusky in #76
• fix(configs): rename message-logger buffer_size → max_entries (22 files) by @cglusky in #79
• feat(rules): adopt action_allowlist + retry-cycle MaxIterations on dev-via-spec rules by @cglusky in #77
• feat(rules): adopt action_allowlist + retry-cycle MaxIterations on research-mode-transition rules by @cglusky in #78
• feat(osh-demo): hybrid model routing per semspec evidence + rate-limit observability by @cglusky in #80
• fix(compose): cap llama.cpp + Rust runtime resources per semspec evidence by @cglusky in #81
• fix(osh-demo): defaults.model must reference an endpoint, not a capability by @cglusky in #82
• docs(adr): split ADR-031 / ADR-033; add ADR-035 + ADR-036 (slice 1/5) by @cglusky in #84
• refactor(slice 2/5): rename vocabulary across personas, configs, code, triples by @cglusky in #85
• feat(slice 3/5): ADR-036 Phase 1 — test-harness manifest seam by @cglusky in #86
• feat(slice 4/5): evidence preprocessor + retire stub block (ADR-036 Phase ...