Skip to main content

Chatbot Sprint 01 — Jira reconciliation pass 08 (2026-07-20)

Jira reconciliation — recheck (2026-07-20, pass 08)​

First fresh pull since pass 6 (pass 7 was same-day registration only). Flow vitals: 67 → 73 issues (+6 net; BOT-4084 pulled out of the sprint, mid-sprint adds incl. 3 fast-track bugs + refine polish); 32 WIP, 31 done (10 new); 22 issues resolved since 2026-07-16 (10 on 07-16, 12 on 07-20 — the sprint's biggest throughput window yet). Oldest in-progress: BOT-3942 ([BE] API webhook result of validation, Eko Aprianto), created 2026-04-07 — 104 days old, a pre-sprint carryover worth a deliberate close-or-descope. Engine ran offline (sprint:reconcile --digest, no local Jira creds); digest pulled via MCP, 73 issues, single page. Code-evidence join: chatbot + chatbot-fe fetched 2026-07-20; qontak-designer, notification-service, hub have no local checkout (all code-evidence verdicts PARTIAL to that extent).

Status of prior findings (by id):

  • F-2 (coverage-gap, AI Agent Impact Report) — STILL OPEN, ⛔ ESCALATED (7th pass). BOT-4568 still has exactly 1 sprint issue (BOT-4645, QA test-case creation, In Testing) and 0 BE/FE tasks. Escalation unchanged: Dimas Fauzi Hidayat decides file-the-tasks vs descope Sprint-1 outcome target 3; risk entry exists (ai-agent-impact-report/delivery/status/, pass 6).
  • F-14 (code-evidence-gap, AI Agent Testing Won't Fix set) — STILL OPEN, ⛔ ESCALATED (6th pass). BOT-3949, BOT-4616, BOT-4617, BOT-4618 still Won't Fix, still no trace on chatbot origin/master (re-grepped after fetch 2026-07-20). If confirmed descoped, testing effort_total_days should shrink — folded into today's ai-agent-testing/delivery/status/2026-07-20.md.
  • F-15 (code-evidence-gap, BOT-4653) — STILL OPEN, ⛔ ESCALATED (6th pass). Still Done, still no trace in chatbot/chatbot-fe (PARTIAL: notification-service unchecked). Owner unchanged (assignee Eko Aprianto / DRI). Re-cited in engineering-q3-2026/delivery/status/2026-07-20.md.
  • F-22 (epic-wiring-drift, BOT-4084 under fast-track Epic BOT-4554) — RESOLVED (left the sprint). BOT-4084 is no longer in sprint 21473, so the sprint-scope question is moot; the BOT-4554 container behaviour continues as new finding F-29.

New findings from the 2026-07-20 pull:

  • F-26 (code-evidence-gap): CRM Deals Custom Fields BOT-4669 (Task 2, AdditionalFieldNormalizer) and BOT-4673 (Task 6, node-registry seed) are board-Done with no commits — and for 4669 no additional_field_normalizer file — on chatbot origin/master (fetched today). Sister tasks BOT-4670/4671/4672 ARE merged (PRs #2394/#2397/#2395), so the gap is specific, not a search-surface artifact. Excluded from the qontak-action progress evidence (1.5 of the 4-of-6 BE md withheld → recorded as 4 md from Tasks 3/4/5 only). Owner: Eko Aprianto — merged elsewhere / superseded / stale board?
  • F-27 (code-evidence-gap): AI Agent Knowledge BOT-4655 / BOT-4656 (adjust knowledge/agent list payloads) Done 2026-07-20 with no trace in chatbot/chatbot-fe. Also off-breakdown (no CONVHIST task rows) → 0 md either way. Owner: Farras Herdicho.
  • F-28 (epic-wiring-drift, orphan — deterministic): BOT-4698 (Failed to create an AI Agent /api/v1/ai_agents/sessions/…, In Review, Agus Suparman) has no parent Epic. Route: jira-alignment-check (board fix — parent it).
  • F-29 (off-plan-load): fast-track Epic BOT-4554 gained 3 mid-sprint bug items — BOT-4690, BOT-4693 (both In Progress, Agus Suparman), BOT-4700 (In Progress, unassigned) — none with a plan queue row, no Expedite-intake record. Real committed load the plan hides, landing on the sprint's busiest BE (Agus also owns the F-30 upgrade set). Owner: Dimas Fauzi Hidayat — route through the adhoc intake gate or record as accepted off-plan load.
  • F-30 (forecast-at-risk / plan-drift, skill-authored): the A1 Ruby/Rails Phase-1 upgrade is being built, not discovered. Plan says "discovery only, build not authorized, excluded from staffing/capacity"; board shows 14 implementation tasks BOT-4623–4636 under bucket Epic BOT-4620 with Tasks 1–7 Ready to Test / In Review / In Progress (Agus Suparman). Also wiring-adjacent: the tasks parent the bucket Epic directly and are not split-linked to Story BOT-4621, so the A1 tracker row still reads backlog while the work runs. Owner: Dimas Fauzi Hidayat — authorize retroactively (update A1 status, plan, capacity) or stop the build. Routed to engineering-q3-2026/delivery/status/2026-07-20.md this pass.
  • status-progress (info, captured this pass — the step-6b computations):
    • Autonomous AI Agent (P1 Refiner): 12.5 / 24.5 md (51%). FE breakdown Tasks 1.1–1.4 done via Story BOT-4574 + Tasks BOT-4579–4586 (Done 2026-07-16), corroborated on chatbot-fe origin/main (PRs #1546/#1548; RefinePanel.vue, RefineOptionCard.vue, useRefineAgent.ts, refine-happy-path.spec.ts all present). Unjoined and NOT counted: BOT-4594 (BE, merged PR #2382, no task token — candidate Tasks 2.3–2.5, up to +5 md, DRI to confirm); BOT-4651 / BOT-4689 (merged polish, no breakdown rows). Join caveat recorded: board Task 1–8 numbering ≠ breakdown 1.1–1.4 ids — content-mapped. Written: dated status + README cache + roadmap row.
    • CRM Deals Custom Fields (qontak-action): 4 of 6 BE breakdown md corroborated done (Tasks 3/4/5 merged; Task 1 Won't Fix; Tasks 2/6 = F-26). README effort_* TBD blocks any progress_pct — evidence parked in the dated status; baseline gap flagged to DRI/PM.
    • AI Agent Testing: count-based only — 7 of 12 board issues done, 3 corroborated (BOT-3941 PR #2393, BOT-4188, BOT-4654 PR #2385); none join to a breakdown task id (BOT-4654 ≈ Task 2.1 is the flagged candidate, 2.5 md). Man-days stay 0 pending DRI joins.
  • Identity/engine corrections: the engine's two roster-mismatch rows are false positives — Muhammad Izzul has BOT-4637 In Progress (roster resolves Izzul correctly; the engine's queue-join missed it — report upstream to sprint-reconcile.mjs), and Dimas Fauzi Hidayat's zero-issue row is the known PM/TPM informational (ex-F-7/F-19). The five umbrella/non-initiative "untracked Epic" false positives from pass 7 (BOT-4602/4609/4482/4500/4662) re-fired unchanged — still not actionable, still worth upstreaming.

This pass does NOT fix: F-2 (with Dimas — file or descope), F-14/F-15/F-26/F-27 (with the named assignees — board/code truth), F-28/F-29 (board fixes via jira-alignment-check / adhoc intake), F-30 (Dimas — authorize or stop; status risk entry written this pass), the BOT-4594 and BOT-4654 join confirmations (DRI), and the qontak-action effort baseline (DRI/PM sizing).

Downstream propagation worklist (this pass)​

#FindingCanonical artifact to reconcileOwning tool / skillState
1Refiner progress 12.5/24.5autonomous-ai-agent/delivery/status/2026-07-20.md → README cache → roadmap rowstep-6b computed, hand-appliedwritten this pass
2QACF BE evidence (4 md)qontak-action/delivery/status/2026-07-20.md (no pct — effort TBD)step-6b + DRI sizing follow-upwritten this pass
3Testing count-note + F-14ai-agent-testing/delivery/status/2026-07-20.mdstep-6b (joins pending DRI)written this pass
4F-30 unauthorized build + F-15engineering-q3-2026/delivery/status/2026-07-20.mdTPM decision → then sprint-planner amend + capacityrisk written; decision open
5Bucket rollup 0-of-4 + roadmap proseroadmap ## Portfolio progress & forecastdeterministicwritten this pass
6F-28 orphan BOT-4698; F-30 tasks not split from BOT-4621Jira boardjira-alignment-checkopen
7F-29 fast-track addsadhoc intake gate / accepted-load noteadhoc-interrupt-doctrine + Dimasopen
8F-2 coverage gapJira board (BE/FE tasks) or plan descopesprint-jira-sync / sprint-planner amendopen (escalated)
9Engine bugs (Izzul queue-join; umbrella false positives)qontak-docs/scripts/sprint-reconcile.mjsupstream issueopen