Remaining Work (7 October 2026)

As of: 2026-09-22 · Release gate: AC01–AC21 (re-baselined), zero critical/high defects, blended cost ≤ $0.15/paid minute, Kevin sign-off.

---

Executive summary

Dimension	Estimate
MVP scope	Auth, positions, candidates, day-window scheduling, credits/Stripe, live AI interview, proctoring, reports, retention, ops
Built	~30–40% — core API/web, ledger, settlement, report generation; not launch-ready
Remaining	~40–50 dev-days open tickets + ~15–25 dev-days FR19, load/capacity, qualification harnesses
Capacity (§18.6)	≤ 270 build-hours; 120 QA-hours final week
Calendar	~15 days to 7 Oct; feature freeze 29 Sep at risk

Working today: Auth through schedule/reserve, candidate lobby/live (Daily), settlement + report Lambda consumers, Stripe credits; no production email, sweepers, report UI/PDF, or FR19. Identity: Firebase in code vs Cognito in spec (T-002).

---

Priority 1: Critical Path

Feature / work item	What it does	Current state	Dependencies	Key files / modules	Risk
Email + ICS invitation system	Candidate day-window invite + .ics; team invite emails	🔴 Not started (dev logger only)	SMTP/SES, WEB_BASE_URL, branding	apps/api/src/modules/interviews/invitations/*, T-017	High
Notification system (async email)	Report-ready/failed, funding alerts, completion, retention warnings	🔴 Not started	T-017 queue + consumer	infra/queues.ts, T-017	High
Interview expiry sweeper	Expire scheduled past expiresAt; release credits; notify	🔴 Not started	Credit release, T-017	interview.model.ts, T-018	High
Credit lot expiry sweeper	Forfeit unreserved credits on expired lots	🟡 Partial (expireLot; no scheduler)	T-009	credit.service.ts, T-018	High
Short-session deterministic result	<60s or <2 substantive answers → insufficient evidence	🟢 Implemented	Bot transcript/substantiveAnswers	report-generation.service.ts	Low
Abuse controls (short-session blocking)	Block wallet gaming on ultra-short sessions	🔴 Not started	Settlement + thresholds	§15	Medium
Proctoring + credibility pipeline (FR19)	Detectors + AI-likeness → credibility score	🔴 Not started (schema only)	G0b; bot + PATCH proctoring	cleverhire-bot, bot-interview.service.ts	High
PDF + evidence export	Recruiter PDF; evidence bundle with expiring URLs	🔴 Not started	Report data, S3	T-015	High
Report viewing UI	Skills, transcript, integrity, recording, export	🔴 Not started	T-014	apps/web, T-015	High
Recording management	Lifecycle, playback URLs, purge	🟡 Partial (Daily enabled; no bucket/lifecycle)	Daily, retention sweeper	daily.service.ts, infra/storage.ts	High
Data retention purge pipeline	Purge at retentionExpiresAt; warnings	🔴 Not started	T-018 + T-017	data-lifecycle.md	High
Reconciliation job (cache vs ledger)	Daily compare; alarm on drift	🔴 Not started	T-009	T-018	High
Capacity scaling (3 → 10 concurrent bots)	Platform 10, 3/workspace; soak + 100-session qual	🔴 Not met (default 3)	Bot max_concurrent_bots	cleverhire-bot/server/services/bot_manager.py	High
Bot E2E production path	Sophia adapter + stable BOT_BASE_URL	🟡 Partial	cleverhire-bot tenant router	bot-interview.service.ts	High
QA contract re-baseline	Align AC/WP with day-window + ICS	🔴 Open	Spec lock v0.5	PRODUCT_CONTEXT.md Q13	High
Wallet qualification (WAL01–WAL24)	≥1k sequences + fault injection	🔴 Not started	Settlement + sweepers	§15.1	High
Proctoring qualification (PRO + held-out set)	Detector metrics, 10× concurrent	🔴 Not started	FR19	§15.2	High
CloudWatch alarms + DLQ ops	DLQ depth, reconciliation drift, stall ≤60s	🔴 Not started	Infra + runbooks	infra/, monitoring.md	Medium
Internal contract docs + adapter tests	Versioned events/API for ATS	🟡 Partial (Zod only)	Stable bot/API	FR18, C14	Medium

Cluster	Tickets / work	Est.
Notifications + ICS	T-017	4d
Sweepers + reconciliation	T-018	3d
Report UI + export	T-015	4d
FR19 proctoring (engineering + qual)	WP25 / not ticketized	10–20d
Capacity 3→10 + soak + 100-session	Load qual	5–8d
WAL + fault injection harness	QA engineering	5–8d
Bot adapter E2E hardening	T-012 remainder	3–5d

---

Priority 2: Should Have

Feature / work item	What it does	Current state	Dependencies	Key files / modules	Risk
Operator console	DLQ replay, org lookup, credit adjustments	🔴 Not started	Auth/RBAC	FR17	Medium
Cost monitoring / alerting	Blended $/minute toward $0.15 gate	🔴 Not measured	Bot costMetrics	nfr-and-quality-gates.md	High
Interview reminder emails	24h before window expiry	🔴 Not started	T-017	FR15	Medium
WAL01–WAL24 automated coverage	Deterministic wallet tests	🔴 Not started	P1 settlement/sweepers	§15.1	High
Chargeback / funding deficit	Chargeback → deficit; block commits	🔴 Not started	Stripe webhooks	credit.service.ts	Medium
Credit restoration flow	WAL13 remedies	🔴 Not started	Operator + ledger	§8	Medium
Resume validation (scanned-only rejection)	Block unreadable résumé at schedule	🟡 Partial	UX + validation	interview.service.ts	Low
Error analytics	Structured error rates, SLOs	🟡 Partial (Pino only)	CloudWatch	observability.ts	Medium
Settings + org branding (T-016)	Admin settings + email branding	🔴 Not started	T-017	T-016	Medium
Disconnect → abandoned sweeper	disconnected >3 min → abandoned + settle	🔴 Not started	T-018	T-018	Medium
Workspace concurrent cap (3)	Per-org live limit	🔴 Not started	Bot/API	FR17	Medium
Provider evidence cleanup (72h)	Deletion attestation	🔴 Not started	Runbooks	§22.4	Medium
On-call + runbooks	DLQ, report retry, rollback	🔴 Not started	Operator console	documentation-gaps.md	Medium
Identity spec alignment	Cognito vs Firebase	🟡 Drift	ADR	T-002, T-003	Medium

---

Priority 3: Can Ship Without

Feature / work item	What it does	Current state	Dependencies	Key files / modules	Risk
Public product / pricing / help / policy pages	Marketing-lite (C03)	🔴 Not started	Static pages	§3.1 item 14	Low
Sample report	Demo without live credits	🔴 Not started	T-015	—	Low
Enhanced PDF styling	Polished export	🔴 Not started	T-015	—	Low
Bulk CSV/XLSX scheduling	100-row import	🔵 OPEN	Q14	C10/FR06	Medium if promised
In-product notification bell	FR15 in-app	🔴 Out of v1	—	—	Low

---

Phase 2 (Post-Launch)

Item	Notes
Dual camera / phone QR companion	G6 / Phase 2
Enterprise SSO (SAML/Okta/Entra)	Excluded Phase 1
ATS integrations + production CleverHire adapter	G5
Apple / Microsoft sign-in	Deferred from C09
Custom domains / white labelling	Out of scope Phase 1
Enhanced pricing tiers	Phase 2
Template library	Dropped
Google / Microsoft calendar	Dropped

---

Dependencies Graph

flowchart TB
  subgraph foundation [Foundation - largely done]
    AUTH[Auth + workspace]
    MODELS[Models + contracts]
    POS[Positions]
    CAND[Candidates + resume]
    CRED[Credits + Stripe]
    SCHED[Interview schedule + reserve]
  end

  subgraph p1_core [P1 critical path]
    EMAIL[T-017 Email + ICS + alerts]
    SWP[T-018 Sweepers + reconciliation]
    BOT[Bot E2E + capacity 10]
    PROC[FR19 Proctoring + credibility]
    RPTGEN[Report generation - done]
    RPTUI[T-015 Report UI + PDF/evidence export]
    RET[Retention purge + recording lifecycle]
    QA[QA re-baseline + WAL/PRO qual]
  end

  subgraph p2 [P2 should have]
    OPS[Operator console + runbooks]
    MON[Cost monitoring + alarms]
    SET[T-016 Settings/branding]
    CB[Chargeback + restoration]
  end

  AUTH --> SCHED
  POS --> SCHED
  CAND --> SCHED
  CRED --> SCHED

  SCHED --> EMAIL
  SCHED --> SWP
  SCHED --> BOT

  BOT --> PROC
  BOT --> RPTGEN
  CRED --> SWP
  SWP --> EMAIL

  RPTGEN --> RPTUI
  PROC --> RPTUI
  SWP --> RET
  RET --> EMAIL

  RPTUI --> QA
  SWP --> QA
  PROC --> QA
  BOT --> QA

  EMAIL --> SET
  SWP --> OPS
  CRED --> CB
  MON --> QA

Critical path: Schedule → Bot E2E → Proctoring → Report UI/export → WAL/PRO qual → Release.

---

Risks & Blockers

#	Risk	Impact	Mitigation status
1	Detector feasibility (FR19)	Release blocker	🔴 G0b; hybrid design not implemented
2	Capacity 3 → 10	AC load gates fail	🔴 max_concurrent_bots=3
3	Cost gate ≤ $0.15/paid minute	Commercial blocker	🔴 Not measured
4	FR18 extraction risks	Data/billing bugs	🟡 Settlement idempotent; rest partial
5	Bot coupling	Dual-product outage	🟡 No versioned adapter
6	Spec / QA drift	False “done”	🔴 Re-baseline OPEN (Q13)
7	Cross-tenant leak	Severity-1	🟡 No AC03 suite
8	Cache vs ledger drift	Billing corruption	🔴 Reconciler not running
9	Stripe / entity / tax	Cannot sell	🟡 External; code exists
10	Report quality / integrity over-claim	Trust/legal	🟡 Qual not run
11	PRODUCT_CONTEXT §25 stale	Planning error	🟡 Use generated docs + this file

Schedule: Feature freeze 29 Sep / G3 6 Oct assume §18.1 complete; ~60–65% work remains — re-baseline milestones (Kevin/Salman).

---

Tickets Reference

Ticket	Title	Priority	Status (2026-09-22)	Remaining work
T-001	Database models + contracts	P0	🟢 Mostly done	v2 fields, proctoring tuning
T-002	Auth infrastructure	P0	🟡 Done (Firebase)	Cognito ADR
T-003	Signup + workspace	P0	🟢 Done	Trial credits config
T-004	Login + recovery	P0	🟢 Done	—
T-005	Team invites	P0	🟡 Partial	Invite email (T-017)
T-006	Positions	P0	🟢 Done	—
T-007	Candidates + resume	P0	🟢 Done	Scanned-only gate
T-008	Notes	P1	🟢 Done	—
T-009	Credits + Stripe	P0	🟡 Mostly done	Chargeback; lot sweeper
T-010	Scheduling + reserve	P0	🟡 Mostly done	Production invitation
T-011	Candidate journey	P0	🟡 Mostly done	FR19 gates; browser QA
T-012	Bot + runtime	P0	🟡 Partial	Production router; proctoring
T-013	Settlement consumer	P0	🟢 Done	WAL fault tests
T-014	Report generation	P0	🟢 Done	—
T-015	Report UI + export	P0	🔴 Not started	Full scope
T-016	Settings + branding	P1	🔴 Not started	API + /settings
T-017	Notifications	P1	🔴 Not started	P1 launch
T-018	Sweepers	P1	🔴 Not started	P1 launch
T-019	App shell	P1	🟢 Done	Settings when T-016
T-020	Interviews CRUD	P0	🟢 Done	Swap ports as needed

Focus order: T-017 → T-018 → Bot E2E + capacity → FR19 MVP → T-015 → WAL/PRO + QA re-baseline → T-016 + monitoring.